Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author.      The effects of productivity and  disturbance on diversity in stream  communities              A thesis presented in partial fulfilment of the requirements for the degree of  Doctor of Philosophy  in Ecology  at  Massey University, Palmerston North, New Zealand.              Jonathan Douglas Tonkin  2010      Abstract  The application of conventional diversity models in streams has had limited success as  stream  communities  often  fit  outside  the model  assumptions.    Of  the  plethora  of  influences  on  stream  macroinvertebrate  diversity,  productivity  and  disturbance  dominate.  Yet there remains to be a consensus on the form of these relationships and  whether productivity and disturbance interact to predict diversity.  Therefore, the aim  of  this  thesis  was  to  the  assess  linkages  between  productivity,  disturbance  and  diversity  in stream macroinvertebrate communities and  further the understanding of  these  links.    I  achieved  this  through multiple  assessments  throughout New  Zealand  (NZ) and Spain between February 2007 and December 2009.  I (i) assessed multivariate  models of productivity‐disturbance‐diversity within NZ streams and examined whether  canopy  presence  modulated  these  links  and  (ii)  whether  land‐use  affected  these  relationships.    I further assessed (iii) the productivity‐diversity relationship  in Spanish  streams  and  whether  this  was  underpinned  by  specific  periphyton‐invertebrate  associations and (iv) if the relationship matched those of NZ and formed a global trend.   Finally, I experimentally examined (v) the influence of spatial scale on productivity‐ and  disturbance‐diversity  relationships.    The  prevailing  premise  throughout  the  thesis  is  that  diversity  increased  with  increasing  productivity  and  declined  with  increasing  disturbance,  although  the  form  of  these  links  varied.    There was  no  evidence  of  a  productivity‐disturbance  interaction  throughout  this  thesis  and  these  relationships  were not affected by land‐use but were by canopy presence.  Where relationships with  productivity  lacked,  I  demonstrate  this  link  may  be  underpinned  by  interactions  between invertebrates and specific forms of algae, as well as simply being a function of  the  range of productivity  assessed or even  the  spatial  scale of  assessment.    In  fact,  productivity setting the upper  limit to richness may be a universal pattern  in streams  with no evidence of productivity  leading  to  greater  competitive  interactions despite  the wide ranges assessed.   Overall, this thesis makes significant progress  in clarifying  these  relationships  in  streams.    I provide  further evidence  suggesting  the DEM does  not  apply  in  stream  communities  and  clearly  demonstrate  additive,  rather  than  interactive, effects of productivity and disturbance.      ii                   To my parents  Roger and Jane Tonkin  who have given me the strength to undertake such a monumental task  and provided never‐ending support.  Thank you.        iii    Acknowledgements  Firstly,  I must  thank my  chief  supervisor  Associate  Professor  Russell  Death  for  his  guidance over the past four years.  Russell has played a large part in my development  as a scientist.   From day one, he encouraged  independent thought and taught me to  stand alone as a scientist.   Russell has shown great patience as a supervisor through  times  of  lack  of motivation  or  even  over‐excitement  at  times.    For  your  help  and  guidance I am deeply grateful.    Thank you  to my  two  co‐supervisors Dr Mike  Joy and Dr Kevin Collier, who  I  knew I could rely on for help and advice throughout my PhD.  I may not have used your  expertise  to  the  fullest  but when  you were  called  on  the  assistance was  extremely  valuable, especially  in the planning stages of the PhD and all the questions along the  way.  I am grateful for the support of Dr Jose Barquin during my trip to Spain and for  arranging much of the logistical side of the research I carried out there.    Thanks  to  the  Ecology Group  staff  including  Fiona Death, Barbara  Just,  Erica  Dahya,  Tracey Harris,  and  Paul  Barrett  for  all  the  logistical  support.    Special  thanks  must go to Cleland Wallace who on several occasions had to help sort supplies or make  things for me at the last minute.  I would like to acknowledge the financial support of a  Massey University Doctoral Scholarship during my study for which I am grateful.    I  am  grateful  to  Keith Wood  at  Ernslaw One  Limited  for  providing  logistical  support and access to Karioi Forest sites.   Dr Michel Dedual and Glenn Mclean of the  Department  of  Conservation,  Turangi  provided  logistical  support  during  the  early  stages  of  the  thesis.   Dr Mike  Scarsbrook  provided  information  on  spring  locations  throughout the North Island.  Dr Barry Biggs of NIWA provided useful help with regard  to  the  design  of  Chapter  2.    Jarrod  Bowler  and  Bonny  Lawrence  at Genesis  energy  provided information on the location of dams within the Tongariro region.    To all my fieldwork assistants: Manas Chakraborty, Logan Brown, Esta Chappell,  Nicki  Atkinson,  Amber  McEwan,  Jess  Costall,  Roger  Tonkin,  Brendon  Fry,  Alana  Lawrence,  Charlie  Charles,  Amy  MacDonald,  Ketan  Parsot  –  thank  you  so  much!   Without  your  help,  the work wouldn’t  have  been  possible.    Special  thanks  goes  to  Brendon Fry for the long days of field work in order to get some fly fishing in.  Thanks  to Alana Lawrence and Travis Delich  for all  the  free accommodation provided during  iv    field work.  Dr Richard Seaton deserves mention for his guidance from a perspective of  someone who had  ‘been  there  and done  that’.    Thanks  also  to  Ecology Group post  graduate  students  for  useful  discussions  on  all  things  ecology,  particularly  Arved  Schwendel who shared supervisor and subject.    I don’t know where to begin to thank my family.  Mum, Dad, Dan and Ami, the  never‐ending and unquestionable support you’ve given me over the past few years  is  nothing short of amazing!  There’s not much else I can say other than I will be forever  grateful and  I hope  this  thesis goes some way towards repaying all  the effort you’ve  put in!    Finally,  no  end  of  thanks will  ever  do  justice  to  the  amount  of  support my  partner, Esta Chappell, has given me.  I’m not sure how you dealt with my daily panic  attacks and mood swings so well, you have the patience of a saint.   You’ve sacrificed  your weekends and social life for far too long.  I for sure would have gotten rid of me  by this stage.  Thank you.                v    Thesis structure and authorship  This PhD  thesis  is a combination of  five  individual papers written as manuscripts  for  publication in refereed journals, in addition to the general introduction and synthesis.   This  has  led  to  unavoidable  repetition  in  some  of  the  content  between  the  five  chapters,  especially  in  the methods  sections.   Also,  the  numbers  for  all  figures  and  tables  restart  for  each  chapter.    Papers  are  co‐authored  in  order  to  acknowledge  significant  contribution  towards  the  paper  including  input  into  the  original  thesis  concept, designing  the  research, administration, and editing manuscripts.   Chapter 2  “Do productivity and disturbance  interact to modulate macroinvertebrate diversity  in  streams?” has been submitted to Oikos and is co‐authored by Russell Death and Kevin  Collier.    Chapter  3  “Consistent  effects  of  productivity  and  disturbance  on  diversity  between  landscapes”  has  been  submitted  to  Ecology  and  is  co‐authored  by  Russell  Death.    Chapter  4  “Productivity‐diversity  relationships  in  streams  and  the  role  of  periphyton  assemblages”  has  not  been  submitted  to  any  journal  but  will  be  co‐ authored  by  Russell  Death  and  Jose  Barquin,  as will  Chapter  5  “Global  patterns  in  stream diversity – productivity relationships”.   Chapter 6 “Scale dependent effects of  productivity and disturbance on diversity” has not been submitted for publication but  will be co‐authored by Russell Death.  For all chapters in this thesis, my input was the  greatest.  I planned the research, carried out the field and lab work, analysed all of the  data and wrote all of the content.      Signed by all involved co‐authors:    Russell Death      Kevin Collier      Jose Barquin    vi    Table of contents  Abstract.....................................................................................................................ii  Dedication.................................................................................................................iii  Acknowledgements...................................................................................................iv  Thesis structure authorship.......................................................................................vi  Chapter 1...................................................................................................................1  General Introduction  Chapter 2.................................................................................................................19  Do productivity and disturbance interact to modulate macroinvertebrate  diversity in streams?  Chapter 3.................................................................................................................51  Consistent effects of productivity and disturbance on diversity between  landscapes.  Chapter 4.................................................................................................................88  Productivity‐diversity relationships in streams and the role of periphyton  assemblages.  Chapter 5...............................................................................................................124  Global patterns in stream diversity – productivity relationships.  Chapter 6...............................................................................................................149  Scale dependent effects of productivity and disturbance on diversity.  Chapter 7...............................................................................................................174  Synthesis  Appendix 1.............................................................................................................181  Raw macroinvertebrate data for Chapter 2  Appendix 2.............................................................................................................196  Site information for Chapter 2  Appendix 3.............................................................................................................198  Raw macroinvertebrate data for Chapter 3  Appendix 4.............................................................................................................216  Raw macroinvertebrate data for Chapter 4  Appendix 5.............................................................................................................222  vii    viii    Raw macroinvertebrate data for Chapter 6  Appendix 6.............................................................................................................225  Co‐authored published paper  Collier, K. J., B. Aldridge, B. J. Hicks, J. Kelly, A. Macdonald, B. J. Smith,  and J. Tonkin. 2009. Ecological values of Hamilton urban streams (North  Island, New Zealand): constraints and opportunities for restoration.  New Zealand Journal of Ecology 33:177‐189.  Appendix 7.............................................................................................................238  Co‐authored published paper  Schwendel, A. C., R. G. Death, I. C. Fuller, and J. D. Tonkin. In press. A  new approach to assess bed stability relevant for invertebrate  communities in upland streams. River Research and Applications.  Chapter 1  General Introduction                      Chapter 1: General Introduction      2    Establishing  what  affects  diversity  in  nature  remains  a  central  theme  in  ecology  (Huston 1994, Rosenzweig 1995, Hubbell 2001).  Although diversity can be affected by  a variety of  factors  (e.g.  temperature, disturbance, productivity, habitat complexity),  there is still considerable debate about how they might interact (Huston 1994, Palmer  1994,  Rosenzweig  1995,  Vinson  and  Hawkins  1998,  Hubbell  2001,  Ricklefs  2004).   Moreover, although  increasing application of multivariate models suggests ecologists  recognise that univariate examinations are limited (Wootton 1998, Mackey and Currie  2001,  Roxburgh  et  al.  2004,  Cardinale  et  al.  2006),  there  remains  considerable  uncertainty as  to which  factors explain  the majority of variation  in diversity  (Mackey  and Currie 2001).  However, disturbance (Connell 1978, Wootton 1998, Lake 2000) and  productivity  (Currie  1991, Waide  et  al.  1999, Mittelbach  et  al.  2001)  continue  to  dominate the research  interests of community ecologists although universal patterns  of diversity remain elusive (Waide et al. 1999, Mackey and Currie 2001, Mittelbach et  al.  2001).    In  fact,  evidence  suggests  that  disturbance  and  productivity  interact  to  affect  diversity  (Sousa  1984,  Pickett  and White  1985,  Currie  1991, Wootton  1998,  Waide et al. 1999, Mittelbach et al. 2001, Death and Zimmermann 2005, Cardinale et  al. 2006).      There have been many attempts to define disturbance such as that put forward  by  Pickett  and White  (1985)  where  they  suggest  disturbance  is  “...  any  relatively  discrete  event  in  time  that  disrupts  ecosystem,  community,  or  population  structure  and changes resources, substrate availability or the physical environment”.  However,  it is tautological to define disturbance by the effects on biota rather than the physical  act  of  disturbance  itself,  especially  with  stream  invertebrate  communities  where  animals are ‘adapted’ to these events (Resh et al. 1988, Lake 2000, Death 2010).  Thus,  using the definitions put forward by Bender et al. (1984) and Glasby and Underwood  (1996),  Lake  (2000)  suggests  disturbance  is  “the  cause  of  a  perturbation”,  where  perturbations  are  a  combination  of  cause  (disturbance)  and  effect  (response  to  disturbance).    Two  of  the  most  longstanding  models  of  diversity  are  the  Intermediate  Disturbance  Hypothesis  (IDH)  (Grime  1973b,  Connell  1978,  Sousa  1979)  and  the  Dynamic Equilibrium Model (DEM) (Huston 1979, 1994).  The IDH predicts diversity to  Chapter 1: General Introduction      3    peak at  intermediate  levels of disturbance, and the DEM builds on this to predict the  position of the peak of diversity to vary with productivity.  Models such as the IDH and  the DEM  assume  a  trade‐off  between  competitive  and  colonisation  ability whereby  organisms are either good colonisers or good competitors, but not both (Chesson and  Huntly 1997, Roxburgh et al. 2004, Cadotte 2007).   However, this does not appear to  occur  in many ecosystems and has  led  some  to  suggest disturbance  simply  removes  taxa  (e.g.  Death  and  Winterbourn  1995,  Death  2002)  as  opposed  to  promoting  diversity at intermediate levels by allowing the representation of the largest number of  life‐history  traits.   Moreover,  the disturbance‐diversity  relationship  is highly  variable  between unimodal, positive  linear and negative  linear  forms, or even no relationship  (Mackey and Currie 2001), and can vary with spatial and temporal scale (Petraitis et al.  1989).   Mackey  and  Currie  (2000)  found  little  evidence  to  suggest  that  disturbance  should play more than a minor role in determining diversity patterns.    The  relationship  between  productivity  and  diversity  is  also  highly  variable  (Currie  1991,  Abrams  1995, Waide  et  al.  1999, Mittelbach  et  al.  2001).    The most  common forms are unimodal (e.g. Grime 1973a, Huston 1979, Rosenzweig 1992, 1995,  Leibold  1999, Waide  et  al.  1999, Mittelbach  et  al.  2001)  or  linear  (e.g.  Currie  and  Paquin 1987, Currie 1991, Abrams 1995, Waide et al. 1999, Gaston 2000, Mittelbach et  al.  2001)  increases  in  diversity  with  increasing  productivity.    The  spatial  scale  of  observation  is  one  of  the major  reasons  for  the  differences  in  the  nature  of  this  relationship such as the switch from unimodal to linear as scale increases (Currie 1991,  Chase and Leibold 2002), but it also differs between ecosystems and organisms (Waide  et  al.  1999, Mittelbach  et  al.  2001).    Small  scale  studies  commonly  find  unimodal  relationships  between  productivity  and  diversity  but  at  greater  spatial  scales  the  pattern is one of monotonic increase in diversity with productivity (Waide et al. 1999,  Mittelbach  et  al.  2001,  Chase  and  Leibold  2002).    The most  commonly  postulated  mechanism  for  the  decline  in  diversity  at  high  productivity  is  competitive  exclusion  (Rosenzweig and Abramsky 1993, Abrams 1995, Waide et al. 1999); although Abrams  (1995) provides several alternative reasons for this pattern such as larger individuals at  higher productivities and  the  fact high productivity environments often occur at  the  extremes of geographical gradients.    Chapter 1: General Introduction      4    Streams and rivers are heavily influenced by disturbance (Resh et al. 1988, Lake  2000,  Death  2008),  but  the  response  of  diversity  to  disturbance  is  likely  to  be  dependent on productivity (Death 2002, Cardinale et al. 2005, Death and Zimmermann  2005,  Cardinale  et  al.  2006).    Several  empirical  studies  have  found  an  interaction  between  productivity  and  disturbance  to  be  important  in  controlling  diversity  (e.g.  Wilson  and  Tilman  2002,  Kneitel  and  Chase  2004,  Death  and  Zimmermann  2005,  Scholes  et  al.  2005).    In  fact,  the  DEM  infers  richness  is  a  product  of  productivity,  disturbance and  their  interaction  (Huston 1979, 1994, Kondoh 2001).   However,  the  highly mobile nature of stream invertebrates  has limited the applicability of common  models of diversity such as the IDH (Grime 1973b, Connell 1978, Sousa 1979) and the  DEM  (Huston  1979,  1994,  Kondoh  2001),  but  see  Townsend  et  al.  (1997).    Streams  have  an  especially  strong  association  between  disturbance  and  productivity,  and  disturbance acts by  removing both  taxa and  their  food  resources  (Lake 2000, Death  2002, Death 2010).    The  requirement  for a  trade‐off  in competitive and colonising ability  in  these  models  (Chesson and Huntly 1997, Roxburgh et al. 2004, Cadotte 2007) and  the  fact  that there appears to be  little evidence for the generality of competitive exclusion  in  stream communities  (Reice 1985, Death and Winterbourn 1995),  indicates  it  is more  likely  that  the  number  of  animals  and/or  taxa  is  determined  by  the  productivity/resource  supply  rates of  a  site  (Death 2002, Gross  and Cardinale 2007,  Death  2008).    Nonetheless,  there  is  certainly  strong  evidence  for  competitive  interactions  and  indeed  competitive  exclusion  in  streams  (McAuliffe  1984,  Kohler  1992), none more evident  than Kohler and Wiley’s  (1997)  investigation of pathogen  outbreaks  in Michigan  streams.    However,  it  appears  physical  processes  are  often  dominant  over  biotic  interactions  in  streams,  especially  in New  Zealand  (Thompson  and  Townsend  2000).    This modulation  of  competition  by  abiotic  and  indeed  biotic  processes  is  well  recognised  for  streams  worldwide making  the  clarification  of  its  effects a challenge (Holomuzki et al. 2010).   An example of this is the strengthening of  predation and competition during low‐flow disturbance events (Dewson et al. 2007).    Accordingly, it is important for benthic ecologists to develop and apply models  relevant  to  the  type  of  communities  present  in  streams.   Death  (2002)  proposed  a  Chapter 1: General Introduction      5    model suggesting disturbance simply resets the colonisation process of  invertebrates;  and productivity, rather than competitive interactions, sets the upper limit to richness  in  streams.    Yet,  there  is  little  consensus  on  quite  how  diversity  responds  to  productivity and disturbance in streams.  In fact, little effort has been placed on clearly  defining the relationship between productivity and diversity in lotic systems.  Some of  the few to look specifically have found both unimodal (Death and Zimmermann 2005)  and  log‐linear  (Death  2002,  Barquin  2004)  increases  in  diversity  with  increasing  productivity.    Streams provide  a unique  testing  ground  for productivity‐disturbance‐ diversity models  because  community  dynamics  are  dominated  by  immigration  and  emigration  rather  than  population  growth  (Downes  1990, Death  2008).    The major  models should not apply  in  lotic communities which are dominated by highly mobile  taxa  rather  than  sedentary  organisms  such  as  plants  and  corals  for which many  of  these models have been developed.    Primary productivity  is typically provided  in streams by periphytic algae which  can vary strongly  in form from prostrate and stalked diatoms through to filamentous  green algae (Allan 1995).  The measurement of primary productivity does not account  for  variation  in  the  form  of  these  producers  which  can  respond  differently  to  environmental  conditions  and  provide  different  habitat  and  resources  for  higher  trophic levels.  When levels of periphyton reach greater densities and epilithic diatoms  are  replaced  by macroalgae  such  as  filamentous  green  algae,  interactions  can  shift  from  simple  plant‐herbivore  interactions  to  more  complex  relationships  where  macroalgae  can both provide and  remove habitat and  food, and  compete  for  space  with certain invertebrate taxa (Dudley et al. 1986).  One example of this is the typical  transition  from  pollution  sensitive  invertebrate  taxa  associated  with  diatom  communities  to  pollution  tolerant  taxa  and  filamentous  algae  growth  forms  often  found  in  summer  low‐flow  conditions  (Suren  et  al.  2003).    Finally,  although  grazing  control  of  periphyton  is  strong  in  lotic  systems  (Hillebrand  2009),  this  can  differ  between different growth  forms  (Feminella and Hawkins 1995).   For example, where  diatom  levels  remain  relatively  stable at  low biomass with  strong  top‐down  control,  filamentous forms can escape grazing and dominate (Suren and Riis 2010).   However,  this may  only  be  an  initial  response  before  grazers  respond  and  regulate  biomass  Chapter 1: General Introduction      6    (Lamberti  et  al.  1989).    Thus,  assessing  the  response  of  higher  trophic  levels  to  periphyton  forms  is  likely  to be complicated by  this differential  top‐down control by  grazers through time and may be a reason for prior research of this relationship being  heavily focused on top‐down, rather than bottom‐up, control.    Measurement  of  both  gross  primary  productivity  (GPP)  and  net  primary  productivity (NPP) in streams has proven troublesome for lotic ecologists compared to  other  systems  (Allan 1995).   There are however,  several approaches  to doing  so, all  with  issues  in  their  reliability  including  biomass  accrual  over  time  and  stream  gas  exchange  (O2  or  CO2)  (Allan  1995),  although  novel  approaches  have  been  taken  in  streams in recent years (Hill and Dimick 2002, Hill et al. 2010).  Often these approaches  are  less  than  satisfactory  for  stream ecosystems with  low productivity  levels due  to  several factors such as the requirement of estimating diffusion between water and air,  and  the  inability  to  limit  these measurements  to  stream  autotrophs.    Due  to  the  difficulties  in measuring primary productivity  in nature, especially  in well spatially or  temporally  replicated  studies,  surrogates  are  widely  used  such  as  biomass,  precipitation, actual evapotranspiration (AET), potential evapotranspiration (PET), and  soil  nutrient  status  (Waide  et  al.  1999, Mittelbach  et  al.  2001, Gillman  and Wright  2006).    Inconsistent  use  of  these  proxies  and  nonlinear  trends with  actual  primary  productivity has stymied attempts at deciphering patterns (Whittaker 2010).    Standing  crop  of  periphyton  biomass  is  commonly  used  as  a  surrogate  for  primary productivity  in streams (Morin et al. 1999).   Morin et al. (1999) reviewed the  relationship between standing crop chlorophyll a and primary productivity  in streams  over  a  large  range of both metrics  spanning  four orders  of magnitude  and  found  a  strong link between the two (r2 = 0.63).  They suggest this surrogate would be useful as  a coarse estimate of production, but also advocate  the addition of  temperature as a  covariate.   However,  it  has  been  recognised  that maximum  biomass  can  be  a  poor  surrogate where herbivory  is severe  (McNaughton 1985).   Another  inherent problem  with this relationship is the decline in primary productivity with increased chlorophyll a  standing  stock  (Morin  et  al.  1999  and  references  therein).    There  are  multiple  mechanisms  that may  lead  to  this  including self shading and shifts  in composition of  periphyton (Dudley et al. 1986, Morin et al. 1999).    Chapter 1: General Introduction      7    It  is also  important  to consider  that primary productivity  is not  the  sole  food  base of stream communities.  Allocthonous detrital input is the other major food base  component of these ecosystems (Minshall 1967, Tank et al. 2010, Finlay 2011).  In fact,  early research suggested streams were heterotrophic in nature, depending heavily on  detritus  inputs  (Tank  et  al.  2010  and  references  therein),  however  recent work  in  tropical  headwater  streams  may  suggest  the  opposite  (Lau  et  al.  2009).    In  New  Zealand, streams often have a large proportion of their length unshaded and thus are  dominated  by  primary  production  (Thompson  and  Townsend  2000)  and  low  proportions of  shredders  (Winterbourn et al. 1981, Winterbourn 1997).   While both  heterotrophic  and  autotrophic  production  are  deemed  to  be  highly  important  to  stream communities,  I have chosen to focus on algal based community production  in  this thesis and accept the caveats with the use of standing crop of periphyton biomass  as a surrogate of primary productivity.    There  is  a  plethora  of  other  factors  that  can  influence  stream  communities  (Vinson  and  Hawkins  1998).    However,  changes  in  land  use  such  as  agricultural  intensification can severely alter communities  living within these  landscapes (Harding  et  al.  1998,  Allan  2004,  Tscharntke  et  al.  2005)  and  are  impacting  negatively  on  habitats worldwide (Tilman 1999, Allan 2004, Foley et al. 2005).   Changes  in  land use  can both alter the composition of stream communities and cause changes in the flood  regime (Rowe et al. 1997, Walsh et al. 2005).   Pastoral development  in particular can  impact  on  stream  communities  through  a  suite  of mechanisms  such  as  changes  to  hydrology,  sediment  regimes,  channel  morphology  and  temperature  (Quinn  2000,  Allan 2004).   However, the effects of agricultural  intensification on flood  influences  is  relatively unknown  (Death 2008).   One  influence  thought  to  strongly  impact  stream  communities which can occur through  land use change  is the presence or absence of  canopy cover.  In small streams this can influence the way in which disturbance affects  macroinvertebrate  diversity  by  altering  primary  productivity  patterns  (Robinson  and  Minshall 1986, Death 2002, Death and Zimmermann 2005, Fuller et al. 2008).  Due to  their strong colonising abilities (Mackay 1992, Allan 1995), post‐flood recovery of the  food  base  in  autotrophic  streams  is  likely  the  major  determinant  of  invertebrate  diversity; however this may not apply in heterotrophic streams if the resource base is  Chapter 1: General Introduction      8    relatively unaffected by disturbance (Death and Zimmermann 2005, Fuller et al. 2008).   This  contrasting  response  of  lotic  invertebrates  between  autotrophic  and  heterotrophic streams is likely to hinder the generalised application of models such as  the DEM.    It  is clear that diversity can be affected by countless environmental  influences  (Rosenzweig  1995).   However,  as with  inconsistent  environmental  sampling,  lack  of  reliable  diversity measures  has  likely  hindered  progress  towards  resolution  of  such  relationships.  Diversity can be coarsely broken down into two main aspects – richness  and  evenness,  and many  indices  attempt  to merge  these  (Magurran  2004).    In  fact  many people,  including myself throughout this thesis,  interchange the terms richness  and diversity (Magurran 2004).  There is still little consensus on which measure is best  as many  can  obscure  patterns,  hence  the  ongoing  development  of  diversity  indices  including measures  of  taxonomic  distinctness  (Clarke  and Warwick  1998).   What  is  more,  these  indices  assess diversity  at  its  lowest  resolution  and do  not  account  for  different  scales  of  diversity  including  β‐diversity  which  has  received  significant  attention in recent years (e.g. Chase and Leibold 2002, Chase 2010).  Moreover, there  is also the question as to whether the number of taxa per unit of area or per unit of  animals  collected, which  can  be  accounted  for  using  rarefaction  (Hurlbert  1971),  is  more  important.   The  approach  I  take  throughout  this  thesis  is  to  assess  whether  richness, namely the number of taxa per unit area, responds to both productivity and  disturbance.  Where other indices have been used it is specific to each chapter.   I take  an approach  that minimises haphazard  selection of multiple  indices  that  is  common  among  researchers  in order  to  find  the best patterns on an ad hoc basis  (Magurran  2004).  It is important to note here that these indices are considered supplementary to  the simple richness measure which forms the basis of this thesis.      A note on statistical paradigms used throughout the thesis  Model  selection  is  a  troublesome  concept  in  ecology.    Historically,  ecologists  have  focused  almost  entirely  on  classical  statistics  revolving  around  the  hypothetico‐ deductive approach with the ever abundant (if not overused) p‐value.  However, this is  complicated  by  the  fact  that  ecological  data  often  does  not  fit  the  assumptions  of  Chapter 1: General Introduction      9    classical  statistics  through  issues  such  as  lack  of  replication,  unequal  variances  and  unbalanced designs.   Burnham and Anderson  (2002) advocate a departure  from  this  with  the  information‐theoretic  approach  focusing  heavily  on model  selection  using  Akaike’s information criterion (AIC) (Akaike 1974) rather than falsifying hypotheses.  It  is  important to note that AIC does not represent a  ‘true’ model, but merely  indicates  the  ‘best’ model of those selected.   Thus,  I take a mix of approaches throughout this  thesis  incorporating mostly  classical hypothesis  testing but when assessing  the  fit of  multiple models I adopt an information‐theoretic approach to model selection.      Thesis structure and aims  These uncertainties  in  relationships between diversity, disturbance and productivity,  especially  in  streams,  led  me  to  this  study  to  examine  the  linkages  between  productivity, disturbance and diversity of macroinvertebrate communities  in streams.   The objectives of each of the remaining six chapters are as follows:    • Chapter  2  examines  the  relationship  between  productivity,  disturbance  and  diversity in mountain streams in the central North Island of New Zealand.  This  chapter assesses the application of the DEM in these stream communities and I  develop a new model for the prediction of taxonomic richness by productivity  and  disturbance  in  streams.    Moreover,  I  assess  whether  the  presence  of  canopy  cover  confounds  the  relationship  between  productivity,  disturbance  and diversity in streams.    • Chapter  3  investigates  whether  the  relationship  between  productivity,  disturbance and diversity is consistent between landscapes.  Due to the level of  land  use  change worldwide  it  is  important  to  consider whether  it  can  affect  these relationships.  Thus, I test whether the productivity‐disturbance‐diversity  relationship differs between a pristine  region  (Tongariro National Park) and a  more altered region (central Hawke’s Bay) through land clearance and pastoral  farming.  The streams in this chapter were sampled multiple times to evaluate  seasonal  variation  and,  to  test  the  productivity‐disturbance‐diversity  relationship, I fit these data to the DEM, the model developed in Chapter 2, and  Death’s (2002) model of diversity.    Chapter 1: General Introduction      10    • Chapter  4  assesses  the  relationship  between  productivity  and  diversity  in  streams  of  Cantabria  in  northern  Spain.    Lack  of  direct  assessment  of  the  productivity‐diversity relationship in streams outside of New Zealand led me to  assess the link in Spanish streams in order to provide more in depth analysis of  this  relationship.    I  assess  whether  there  is  a  clear  relationship  between  productivity  and  diversity  in  these  streams,  or whether  underlying  biological  interactions between  invertebrates and periphyton are more  important.   That  is, I test whether underlying relationships between invertebrates and the main  growth  forms  of  periphyton  are  better  at  predicting  community metrics  in  stream than a simple productivity measure.    • Chapter  5  examines  whether  there  is  a  consistent  productivity‐diversity  relationship  in  lotic  systems  at  two  locations  around  the  world.    For  this  I  compare the productivity‐diversity relationship between streams of the central  North Island of New Zealand and those of northern Spain.  I further explore this  relationship  to  assess  whether  differences  in  patterns  are  simply  due  to  measurement  along  different  environmental  scales  and  assess  whether  combining these regions can yield a universal productivity‐diversity relationship  for streams.    • Chapter 6 uses experimental channels  in spring‐fed streams  in order to assess  both  productivity‐  and  disturbance‐diversity  relationships  and whether  these  differ with  the  scale of observation.    Specifically,  I use proxies of  large  scale  patterns such as α, β and γ diversity by assessing taxonomic richness at within‐  and between‐stream scales, and between‐stream community dissimilarity; and  I examine whether the form of these relationships differ between scales using  experimental  levels of productivity and disturbance.    I discuss possible causes  for differences between scales such as community assembly rules.    • Chapter 7 briefly synthesises the findings throughout the five main chapters of  this thesis.          Chapter 1: General Introduction      References  11    Abrams,  P. A.  1995. Monotonic  or  unimodal  diversity  productivity  gradients  ‐ What  does competition theory predict. Ecology 76:2019‐2027.  Akaike,  H.  1974.  New  look  at  statistical‐model  identification.  Ieee  Transactions  on  Automatic Control AC19:716‐723.  Allan, J. D. 1995. Stream Ecology: Structure and Function of Running Waters. Chapman  and Hall, London.  Allan,  J. D.  2004.  Landscapes  and  riverscapes:  the  influence  of  land  use  on  stream  ecosystems. Annual Review of Ecological and Evolutionary Systematics 35:257‐ 284.  Barquin, J. 2004. Spatial patterns of invertebrate communities in spring and runoff‐fed  streams.  PhD  thesis.  PhD  thesis, Massey  University,  Palmerston  North,  New  Zealand.  Bender,  E.  A.,  T.  J.  Case,  and  M.  E.  Gilpin.  1984.  Perturbation  Experiments  in  Community Ecology ‐ Theory and Practice. Ecology 65:1‐13.  Burnhamn,  K.  P.,  and  D.  R.  Anderson.  2002.  Model  Selection  and  Multi‐model  inference:  A  practical  Information‐theoretic  Approach.  Springer,  New  York,  USA.  Cadotte, M. W. 2007. Competition‐colonization  trade‐offs and disturbance effects at  multiple scales. Ecology 88:823‐829.  Cardinale, B. J., H. Hillebrand, and D. F. Charles. 2006. Geographic patterns of diversity  in  streams  are  predicted  by  a  multivariate  model  of  disturbance  and  productivity. Journal of Ecology 94:609‐618.  Cardinale, B. J., M. A. Palmer, A. R. Ives, and S. S. Brooks. 2005. Diversity‐productivity  relationships  in streams vary as a  function of  the natural disturbance  regime.  Ecology 86:716‐726.  Chase, J. M. 2010. Stochastic Community Assembly Causes Higher Biodiversity in More  Productive Environments. Science 328:1388‐1391.  Chase,  J.  M.,  and  M.  A.  Leibold.  2002.  Spatial  scale  dictates  the  productivity‐ biodiversity relationship. Nature 416:427‐430.  Chapter 1: General Introduction      12    Chesson, P., and N. Huntly. 1997. The roles of harsh and fluctuating conditions  in the  dynamics of ecological communities. The American Naturalist 150:519‐553.  Clarke,  K.  R.,  and  R.  M.  Warwick.  1998.  A  taxonomic  distinctness  index  and  its  statistical properties. Journal of Applied Ecology 35:523‐531.  Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs. Science 199:1302‐ 1310.  Currie, D. J. 1991. Energy and large‐scale patterns of animal‐species and plant‐species  richness. American Naturalist 137:27‐49.  Currie,  D.  J.,  and  V.  Paquin.  1987.  Large‐scale  biogeographical  patterns  of  species  richness of trees. Nature 329:326‐327.  Death, R. G. 2002. Predicting invertebrate diversity from disturbance regimes in forest  streams. Oikos 97:18‐30.  Death, R. G. 2008. Effects of floods on aquatic invertebrate communities.in J. Lancaster  and R. A. Briers, editors. Insects: Challenges to Populations. CAB International,  UK, pp 103‐121.  Death,  R.  G.  2010.  Disturbance  and  riverine  benthic  communities:  What  has  it  contributed  to  general  ecological  theory?  River  Research  and  Applications  26:15‐25.  Death,  R.  G.,  and  M.  J.  Winterbourn.  1995.  Diversity  patterns  in  stream  benthic  invertebrate communities:  the  influence of habitat  stability. Ecology 76:1446‐ 1460.  Death,  R.  G.,  and  E. M.  Zimmermann.  2005.  Interaction  between  disturbance  and  primary  productivity  in  determining  stream  invertebrate  diversity.  Oikos  111:392‐402.  Dewson, Z. S., A. B. W. James, and R. G. Death. 2007. A review of the consequences of  decreased  flow  for  instream  habitat  and macroinvertebrates.  Journal  of  the  North American Benthological Society 26:401‐415.  Downes,  B.  J.  1990.  Patch  dynamics  and  mobility  of  fauna  in  streams  and  other  habitats. Oikos 59:411‐413.  Chapter 1: General Introduction      13    Dudley, T. L., S. D. Cooper, and N. Hemphill. 1986. Effects of Macroalgae on a Stream  Invertebrate Community. Journal of the North American Benthological Society  5:93‐106.  Feminella, J. W., and C. P. Hawkins. 1995. Interactions between stream herbivores and  periphyton: A quantitative analysis of past experiments.  Journal of  the North  American Benthological Society 14:465‐509.  Finlay, J. C. 2011. Stream size and human influences on ecosystem production in river  networks. Ecosphere 2:art87.  Foley, J. A., R. DeFries, G. P. Asner, C. Barford, G. Bonan, S. R. Carpenter, F. S. Chapin,  M. T. Coe, G. C. Daily, H. K. Gibbs, J. H. Helkowski, T. Holloway, E. A. Howard, C.  J.  Kucharik,  C. Monfreda,  J. A.  Patz,  I.  C.  Prentice, N.  Ramankutty,  and  P.  K.  Snyder. 2005. Global consequences of land use. Science 309:570‐574.  Fuller, R. L., C. LaFave, M. Anastasi, J. Molina, H. Salcedo, and S. Ward. 2008. The role  of  canopy  cover  on  the  recovery  of  periphyton  and  macroinvertebrate  communities after a month‐long flood. Hydrobiologia 598:47‐57.  Gaston, K. J. 2000. Global patterns in biodiversity. Nature 405:220‐227.  Gillman,  L. N.,  and  S. D. Wright.  2006.  The  influence  of  productivity  on  the  species  richness of plants: A critical assessment. Ecology 87:1234‐1243.  Glasby, T. M., and A. J. Underwood. 1996. Sampling to differentiate between pulse and  press perturbations. Environmental Monitoring and Assessment 42:241‐252.  Grime,  J. P. 1973a. Competitive exclusion  in herbaceous vegetation. Nature 242:344‐ 347.  Grime,  J.  P.  1973b.  Control  of  species  density  in  herbaceous  vegation.  Journal  of  Environmental Management 1:151‐167.  Gross, K., and B. J. Cardinale. 2007. Does species richness drive community production  or  vice  versa?  Reconciling  historical  and  contemporary  paradigms  in  competitive communities. American Naturalist 170:207‐220.  Harding,  J.  S.,  E.  F.  Benfield,  P.  V.  Bolstad,  and  E.  B.  D.  Jones  III.  1998.  Stream  biodiversity: The ghost of  land use past. Proceedings of the National Academy  of Sciences 95:14843‐14847.  Chapter 1: General Introduction      14    Hill, W. R.,  and  S. M. Dimick.  2002.  Effects  of  riparian  leaf  dynamics  on  periphyton  photosynthesis  and  light  utilisation  efficiency.  Freshwater  Biology  47:1245‐ 1256.  Hill,  W.  R.,  J.  Rinchard,  and  S.  Czesny.  2010.  Light,  nutrients  and  the  fatty  acid  composition of stream periphyton. Freshwater Biology 56:1825‐1836.  Hillebrand,  H.  2009. Meta‐analysis  of  grazer  control  of  periphyton  biomass  across  aquatic ecosystems. Journal of Phycology 45:798‐806.  Holomuzki,  J.  R.,  J.  W.  Feminella,  and  M.  E.  Power.  2010.  Biotic  interactions  in  freshwater  benthic  habitats.  Journal  of  the  North  American  Benthological  Society 29:220‐244.  Hubbell,  S.  P.  2001.  The  unified  neutral  theory  of  biodiversity  and  biogeography.  Princeton University Press.  Hurlbert, S. H. 1971. The non‐concept of  species diversity: a critique and alternative  parameters. Ecology 52:577‐586.  Huston, M. 1979. A general hypothesis of  species diversity. The American Naturalist  113:81‐100.  Huston,  M.  1994.  Biological  Diversity:  The  Coexistence  of  Species  on  Changing  Landscapes. Cambridge University Press, Cambridge, UK.  Kneitel, J. M., and J. M. Chase. 2004. Disturbance, predator, and resource interactions  alter container community composition. Ecology 85:2088‐2093.  Kohler,  S.  L.  1992.  Competition  and  the  structure  of  a  benthic  stream  community.  Ecological Monographs 62:165‐188.  Kohler, S. L., and M.  J. Wiley. 1997. Pathogen outbreaks  reveal  large‐scale effects of  competition in stream communities. Ecology 78:2164‐2176.  Kondoh, M.  2001. Unifying  the  relationships  of  species  richness  to  productivity  and  disturbance.  Proceedings  of  the  Royal  Society  of  London  Series  B‐Biological  Sciences 268:269‐271.  Lake, P. S. 2000. Disturbance, patchiness, and diversity in streams. Journal of the North  American Benthological Society 19:573‐592.  Chapter 1: General Introduction      15    Lamberti, G. A., S. V. Gregory, L. R. Ashkenas, A. D. Steinman, and C. D. McIntire. 1989.  Productive  capacity  of  periphyton  as  a  determinant  of  plant  herbivore  interactions in streams. Ecology 70:1840‐1856.  Lau, D. C. P., K. M. Y. Leung, and D. Dudgeon. 2009. Are autochthonous  foods more  important  than  allochthonous  resources  to  benthic  consumers  in  tropical  headwater  streams?  Journal  of  the  North  American  Benthological  Society  28:426‐439.  Leibold,  M.  A.  1999.  Biodiversity  and  nutrient  enrichment  in  pond  plankton  communities. Evolutionary Ecology Research 1:73‐95.  Mackay, R. J. 1992. Colonization by Lotic Macroinvertebrates  ‐ a Review of Processes  and Patterns. Canadian Journal of Fisheries and Aquatic Sciences 49:617‐628.  Mackey,  R.  L.,  and  D.  J.  Currie.  2000.  A  re‐examination  of  the  expected  effects  of  disturbance on diversity. Oikos 88:483‐493.  Mackey,  R.  L.,  and  D.  J.  Currie.  2001.  The  diversity‐disturbance  relationship:  Is  it  generally strong and peaked? Ecology 82:3479‐3492.  Magurran, A. E. 2004. Measuring Biological Diversity. Blackwell Science  Ltd., Oxford,  UK.  McAuliffe,  J.  R.  1984.  Competition  for  space,  disturbance,  and  the  structure  of  a  benthic stream community. Ecology 65:894‐908.  McNaughton,  S.  J.  1985.  Ecology  of  a  grazing  ecosystem  ‐  The  Serengeti.  Ecological  Monographs 55:259‐294.  Minshall,  G. W.  1967.  Role  of  allochthonous  detritus  in  the  trophic  structure  of  a  woodland springbrook community. Ecology 48:139‐149.  Mittelbach, G. G., C. F. Steiner, S. M. Scheiner, K. L. Gross, H. L. Reynolds, R. B. Waide,  M. R. Willig, S. I. Dodson, and L. Gough. 2001. What is the observed relationship  between species richness and productivity? Ecology 82:2381‐2396.  Morin, A., W. Lamourex, and  J. Busnarda. 1999. Empirical models predicting primary  productivity  from chlorophyll a and water temperature  for stream periphyton  and  lake  and  ocean  phytoplankton.  Journal  of  the  North  American  Benthological Society 18:299‐307.  Chapter 1: General Introduction      16    Palmer,  M.  W.  1994.  Variation  in  species  richness  ‐  towards  a  unification  of  hypotheses. Folia Geobotanica & Phytotaxonomica 29:511‐530.  Petraitis, P. S., R. E. Latham, and R. A. Niesenbaum. 1989. The Maintenance of Species‐ Diversity by Disturbance. Quarterly Review of Biology 64:393‐418.  Pickett, S. T. A., and P. S. White. 1985. The ecology of natural disturbance and patch  dynamics. Academic Press.  Quinn, J. M. 2000. Effects of pastoral development. Pages 208‐229  in K. J. Collier and  M.  J. Winterbourn,  editors.  New  Zealand  stream  invertebrates:  ecology  and  implications for management. The Caxton Press, Christchurch, New Zealand.  Reice, S. R. 1985. Experimental disturbance and the maintenance of species‐diversity in  a stream community. Oecologia 67:90‐97.  Resh, V. H., A. V. Brown, A. P. Covich, M. E. Gurtz, H. W. Li, G. W. Minshall, S. R. Reice,  A. L. Sheldon, J. B. Wallace, and R. C. Wissmar. 1988. The role of disturbance in  stream  ecology.  Journal  of  the North  American  Benthological  Society  7:433‐ 455.  Ricklefs, R.  E.  2004. A  comprehensive  framework  for  global patterns  in biodiversity.  Ecology Letters 7:1‐15.  Robinson, C. T., and G. W. Minshall. 1986. Effects of disturbance frequency on stream  benthic community structure in relation to canopy cover and season. Journal of  the North American Benthological Society 5:237‐248.  Rosenzweig, M. L. 1992. Species‐diversity gradients ‐ we know more and less than we  thought. Journal of Mammalogy 73:715‐730.  Rosenzweig, M.  L.  1995.  Species Diversity  in  Space  and  Time.  Cambridge University  Press, Cambridge, UK.  Rosenzweig, M. L., and Z. Abramsky. 1993. How are Diversity and Productivity Related?  Pages  52‐65  in  R.  E.  Ricklefs  and  D.  Schluter,  editors.  Species  Diversity  in  Biological Communities. University of Chicago Press, Chicago, Illinois, USA.  Rowe, L., B. Fahey, R. Jackson, and M. Duncan. 1997. Effects of land use on floods and  low flows. Pages 89‐102 in M. P. Mosely and C. P. Pearson, editors. Floods and  Droughts: the New Zealand Experience. New Zealand Hydrological Society Inc.,  Christchurch, New Zealand.  Chapter 1: General Introduction      17    Roxburgh,  S.  H.,  K.  Shea,  and  J.  B.  Wilson.  2004.  The  intermediate  disturbance  hypothesis:  Patch  dynamics  and mechanisms  of  species  coexistence.  Ecology  85:359‐371.  Scholes, L., P. H. Warren, and A. P. Beckerman. 2005. The combined effects of energy  and  disturbance  on  species  richness  in  protist  microcosms.  Ecology  Letters  8:730‐738.  Sousa, W. P. 1979. Disturbance in marine intertidal boulder fields: the nonequilibrium  maintenance of species diversity. Ecology 60:1225‐1239.  Sousa, W. P. 1984. The role of disturbance  in natural communities. Annual Review of  Ecology and Systematics 15:353‐391.  Suren, A. M., B.  J.  F. Biggs, M.  J. Duncan,  L. Bergey,  and  P.  Lambert.  2003. Benthic  community  dynamics  during  summer  low‐flows  in  two  rivers  of  contrasting  enrichment 2.  Invertebrates. New  Zealand  Journal of Marine  and  Freshwater  Research 37:71‐83.  Suren, A. M.,  and  T. Riis.  2010.  The  effects of plant  growth on  stream  invertebrate  communities  during  low  flow:  a  conceptual  model.  Journal  of  the  North  American Benthological Society 29:711‐724.  Tank, J. L., E. J. Rosi‐Marshall, N. A. Griffiths, S. A. Entrekin, and M. L. Stephen. 2010. A  review of allochthonous organic matter dynamics and metabolism  in streams.  Journal of the North American Benthological Society 29:118‐146.  Thompson,  R. M.,  and  C.  R.  Townsend.  2000.  New  Zealand's  stream  invertebrate  communities:  an  international  perspective.in  K.  J.  Collier  and  M.  J.  Winterbourn,  editors.  New  Zealand  stream  invertebrates:  ecology  and  implications for management. New Zealand Limnological Society, Hamilton.  Tilman, D. 1999. Global environmental impacts of agricultural expansion: The need for  sustainable  and  efficient  practices.  Proceedings  of  the  National  Academy  of  Sciences of the United States of America 96:5995‐6000.  Townsend, C. R., M. R. Scarsbrook, and S. Doledec. 1997. The intermediate disturbance  hypothesis, refugia, and biodiversity  in streams. Limnology and Oceanography  42:938‐949.  Chapter 1: General Introduction      18    Tscharntke,  T.,  A.  M.  Klein,  A.  Kruess,  I.  Steffan‐Dewenter,  and  C.  Thies.  2005.  Landscape  perspectives  on  agricultural  intensification  and  biodiversity  ‐  ecosystem service management. Ecology Letters 8:857‐874.  Vinson, M. R., and C. P. Hawkins. 1998. Biodiversity of stream insects: variation at local,  basin, and regional scales. Annual Review of Entomology 43:271‐293.  Waide, R. B., M. R. Willig, C.  F.  Steiner, G. Mittelbach,  L. Gough,  S.  I. Dodson, G. P.  Juday,  and  R.  Parmenter.  1999.  The  relationship  between  productivity  and  species richness. Annual Review of Ecology and Systematics 30:257‐300.  Walsh, C.  J., A. H. Roy,  J. W. Feminella, P. D. Cottingham, P. M. Groffman, and R. P.  Morgan. 2005. The urban stream syndrome: current knowledge and the search  for a cure. Journal of the North American Benthological Society 24:706‐723.  Whittaker,  R.  J.  2010.  In  the  dragon's  den:  a  response  to  the meta‐analysis  forum  contributions. Ecology 91:2568‐2571.  Wilson,  S. D.,  and D.  Tilman.  2002. Quadratic  variation  in  old‐field  species  richness  along gradients of disturbance and nitrogen. Ecology 83:492‐504.  Winterbourn, M.  J.  1997.  New  Zealand  mountain  stream  communities:  stable  yet  disturbed?  Pages  31‐54  in  B.  Streit,  T.  Stadler,  and  C.  M.  Lively,  editors.  Evolutionary  Ecology  of  Freshwater  Animals.  Birkhauser  Verlag,  Basel,  Switzerland.  Winterbourn,  M.  J.,  J.  S.  Rounick,  and  B.  Cowie.  1981.  Are  New  Zealand  stream  ecosystems  really  different? New  Zealand  Journal  of Marine  and  Freshwater  Research 15:321‐328.  Wootton,  J.  T.  1998.  Effects  of  disturbance  on  species  diversity:  A  multitrophic  perspective. American Naturalist 152:803‐825.      Chapter 2  Do productivity and disturbance interact to  modulate macroinvertebrate diversity in  streams?                          Chapter 2: Interactive effects of productivity and disturbance on stream diversity      Abstract  20    The application of diversity models  in  lotic systems has been somewhat unsuccessful  as stream communities don’t often fit conventional model assumptions.  Although it is  clear  that  disturbance  and  productivity  are  dominant  forces  in  lotic  systems,  the  inability  of  the  Dynamic  Equilibrium  Model  (DEM)  to  explain  observed  patterns  suggests  specific models are  required  for  these  systems  that are not  constrained by  competitive‐colonising trade‐offs.  We set out to (i) test the application of the DEM in  47 mountain  streams  in  the  central North  Island of New  Zealand  and  (ii) develop  a  modified model based on observed  relationships  in  this  study  that  accounts  for  the  confounding  influence of canopy cover.   Although canopy cover had no  influence on  productivity,  a  link  between  disturbance  and  productivity  was  only  found  in  open  canopy  streams  where  taxonomic  richness  increased  log‐linearly  with  increasing  productivity and peaked at  intermediate  levels of disturbance.   When accounting  for  evenness of communities, the results were more inconsistent exhibiting a decline with  disturbance,  but  not  at  open  canopy  sites.   However,  both  taxonomic  richness  and  Simpson’s diversity index were higher in closed canopy streams.  Overall, we show that  richness of stream invertebrates in the present study is a product of both disturbance  and  productivity  but  not  of  their  interactive  effects.   Although  there was  a  peak  in  richness at intermediate rates of disturbance, our results do not support the DEM.  We  suggest the combined effects of productivity and disturbance are additive rather than  multiplicative supporting other studies that suggest productivity simply sets the upper  limit to richness in streams.                   Chapter 2: Interactive effects of productivity and disturbance on stream diversity      Introduction  21    Establishing what factors control diversity in nature has long been an important theme  of research in ecology (Huston 1994, Rosenzweig 1995, Hubbell 2001).  Although many  factors  can  affect  diversity  (Palmer  1994,  Vinson  and  Hawkins  1998),  there  is  still  considerable  debate  over  how  they  might  interact  (Hubbell  2001).    In  particular,  several  studies  in  a  variety  of  ecosystems  have  demonstrated  that  disturbance  and  productivity  can  interact  to  affect  diversity,  but  the  form  of  this  interaction  varies  (Sousa 1984, Pickett and White 1985, Currie 1991, Wootton 1998, Waide et al. 1999,  Mittelbach et al. 2001, Death and Zimmermann 2005, Cardinale et al. 2006).    Disturbance  is one of the major structuring forces  in  lotic systems (Resh et al.  1988,  Lake  2000,  Death  2010).    A  dominant  paradigm  in  disturbance  ecology,  the  Intermediate Disturbance Hypothesis (IDH), has been amongst the most widely applied  ecological  theories  (Grime  1973b,  Connell  1978,  Sousa  1979).    However,  there  has  been  little  empirical  support  for  this  hypothesis  in  streams,  possibly  because many  stream  organisms  are  highly mobile  (but  see  Townsend  et  al.  1997).    In  fact,  the  literature  suggests  that  the  disturbance‐diversity  relationship  is  highly  variable  in  nature  generally  (Mackey  and  Currie  2001).   What  is more  surprising  is  that many  studies  show no or weak effects of disturbance on diversity or community  structure  (Mackey and Currie 2000).  In lotic systems, the disturbance‐diversity relationship can  be  confounded  by  the  fact  that  disturbance  not  only  acts  directly  on  benthic  invertebrates  but  indirectly  by  the  removal  of  food  resources  (Death  2002).    Thus  several  authors  have  proposed  that  the  nature  of  the  disturbance‐diversity  relationship  is  controlled  by  habitat  productivity  altering  population  growth  rates  (Huston 1979, 1994, Kondoh 2001, Cardinale et al. 2006).    The relationship between productivity and diversity has also been an important  research theme in ecology (Currie 1991, Abrams 1995, Waide et al. 1999, Mittelbach et  al. 2001).  However, just as with the disturbance‐diversity relationship, the form of the  relationship can be quite variable (Abrams 1995, Waide et al. 1999, Mittelbach et al.  2001).   The most  commonly  reported  relationships are unimodal  (e.g. Grime 1973a,  Huston 1979, Rosenzweig 1992, 1995, Leibold 1999, Waide et al. 1999, Mittelbach et  Chapter 2: Interactive effects of productivity and disturbance on stream diversity      22    al. 2001) or linear (e.g. Currie and Paquin 1987, Currie 1991, Abrams 1995, Waide et al.  1999,  Gaston  2000,  Mittelbach  et  al.  2001)  increases  in  diversity  with  increasing  productivity.    The  variation  in  observed  patterns  may  be  a  result  of  the  scale  of  observation which  has  ranged  from  local,  to  regional  and  global  comparisons  (e.g.  Currie  1991,  Chase  and  Leibold  2002).    However,  local  scale  studies  often  find  unimodal relationships  (Waide et al. 1999, Mittelbach et al. 2001, Chase and Leibold  2002)  which  are  potentially  explained  by  many  mechanisms  (Rosenzweig  and  Abramsky  1993, Abrams  1995, Waide  et  al.  1999)  that  often  require  some  form  of  competitive trade‐off.    An extension of the IDH, the dynamic equilibrium model (DEM) (Huston 1979,  1994), predicts that the level of disturbance maximising diversity changes with habitat  productivity.  Using the patch occupancy models of Hastings (1980) and Tilman (1994),  Kondoh (2001) expanded this model to account for metapopulation dynamics, multiple  trophic levels and patchy disturbances.  This modified model provides an alternative to  prior models by allowing disturbances to create niche opportunities for the expression  of various life‐history traits.  However, like the DEM, it also predicts that diversity will  peak at  intermediate  levels of productivity and disturbance with  the position of  the  peak on one scale depending on the level of the other.  Models such as the IDH, DEM  and Kondoh’s model assume a trade‐off between competitive and colonisation ability  whereby  organisms  are  either  good  colonisers  or  good  competitors  but  not  both  (Chesson and Huntly 1997, Roxburgh et al. 2004, Cadotte 2007).  There appears to be  little  evidence  for  the  generality  of  competitive  exclusion  in  stream  communities  (Reice 1985, Death and Winterbourn 1995), and  it appears more  likely  that  carrying  capacity may be determined by the productivity/resource supply rates of a site (Gross  and  Cardinale  2007,  Death  2008).    Accordingly,  Death  (2002)  proposed  that  in  the  absence  of  disturbance,  resource  levels  rather  than  competitive  interactions  set  an  upper  limit  to  the  richness  of  communities  through  colonisation, whilst  disturbance  resets  the  colonisation  process  by  removing  animals  and  thus  taxa,  and  resetting  resource levels.    The presence or absence of  canopy  cover  in  small  streams  can  influence  the  way  in which  disturbance  affects macroinvertebrate  diversity  by  regulating  primary  Chapter 2: Interactive effects of productivity and disturbance on stream diversity      23    productivity (Robinson and Minshall 1986, Death 2002, Death and Zimmermann 2005,  Fuller et al. 2008).   As a  large proportion of  lotic  invertebrates have good colonising  abilities  after  disturbance  (Mackay  1992,  Allan  1995),  it  is  likely  that  post‐flood  recovery  of  the  food  base  is  the  major  determinant  of  invertebrate  diversity  in  autotrophic  streams  (Death  and  Zimmermann  2005).    This may  not  be  the  case  in  heterotrophic streams if the resource base is relatively unaffected by disturbance (i.e.  as much organic matter is washed in as is washed out).  Fuller et al. (2008) suggest that  when  periphyton  recovery  is  not  limited  by  nutrients,  open  canopy  sites  are more  resilient  to disturbance  than  sites with  canopy  cover.   Conversely,  sites with  canopy  cover  are  more  resistant  because  the  invertebrate  community  recovery  is  not  as  dependent on periphyton re‐growth post disturbance as in open canopy sites (Fuller et  al.  2008).    The  contrasting  response  of  lotic  invertebrates  to  productivity  and  disturbance  between  autotrophic  and  heterotrophic  streams  is  likely  to  hinder  the  generalised application of models such as the DEM.  That is, where retention of organic  matter  in  forested streams  is sufficient, macroinvertebrate communities are  likely  to  respond  differently  to  disturbance  than  in  autotrophic  streams  as  they  are  less  dependent on primary productivity.    Several studies have recently assessed the DEM in a variety of ecosystems (e.g.  Scholes et al. 2005, Cardinale et al. 2006, Svensson et al. 2007, Haddad et al. 2008,  Sugden et al. 2008).   However, the results have been equivocal with the response of  communities  to  productivity  and  disturbance  varying  inconsistently  between  ecosystems and little evidence of interactive effects (but see Cardinale et al. 2006).  In  this  study  we  test  the  DEM  on  benthic  invertebrate  communities  from  mountain  streams  in  the  central  North  Island  of  New  Zealand.   We  investigate whether  the  observed  levels  of  productivity  and  disturbance  are  sufficient  to  explain  diversity  patterns exclusively in these streams, or whether the relationship is modulated by the  presence of canopy cover over the stream.  We hypothesise that benthic invertebrate  diversity  is  a product of  the  interaction between  substrate disturbance  and primary  productivity, assessed as bed stability and periphyton biomass, respectively, and that  this relationship will be stronger at open canopy sites than at sites with canopy due to  tighter coupling with algal  food  resources.   We discuss whether diversity patterns  in  Chapter 2: Interactive effects of productivity and disturbance on stream diversity      24    these streams can be better explained by a modified productivity‐disturbance‐diversity  model that is not constrained by the competitive‐colonising trade‐off inherent with the  DEM.       Methods  Study sites  Forty seven first‐ to sixth‐order streams and rivers were selected for sampling around  the mountains  of  the  Tongariro  National  Park,  New  Zealand  (Fig.  1).    The  park  is  dominated by the central volcanic massif of Mt Ruapehu (2,797 m asl), Mt Ngauruhoe  (2,287 m asl) and Mt Tongariro (1,967 m asl) and the Tihia‐kakaramea volcanic massif  to the north made up of predominantly andesitic geology.  The Kaimanawa Ranges rise  to ~1,799 m asl to the east of the Tongariro National Park and are made up of Torlesse  Group graywackes and argillites, with variable coverings of volcanic ash deposits.  The  northern and western parts of the park have an average rainfall of 1,800‐3,500 mm yr‐ 1, with the south and east only receiving around 1,100 mm yr‐1 due to the rain‐shadow  cast by the three mountains from the prevailing westerly winds.  Vegetation within and  around the park varies from broadleaf‐podocarp, mixed beech‐podocarp, exotic Pinus  radiata plantation, native  tussock and scrubland,  to bare ground  in  the eastern rain‐ shadow of the three central volcanoes.    All sampling sites had  less than 10 % catchment pastoral  land use and greater  than  90 %  volcanic hard  sedimentary  geology.    Thus, water quality  at  these  sites  is  relatively unimpaired by human  influences, and  flows are unmodified other than the  effects  of  run‐of‐river  hydro‐electric  dams  at  ten  sites.    Elsewhere  sites  varied  hydrologically from stable spring‐fed streams, to runoff‐fed streams.      Biological collections  Macroinvertebrates were sampled on one occasion  from early February  to  late April  2007.  Five 0.1 m2 Surber samples (500 µm mesh) were taken from random locations in  riffles throughout approximately 50 m reaches at each site.   Samples were preserved  in 10 % formalin and later identified in the laboratory to the lowest possible taxonomic  Chapter 2: Interactive effects of productivity and disturbance on stream diversity      25    level using available keys (e.g. Towns and Peters 1996, Winterbourn et al. 2000).  Taxa  that  could not be  taken  to  species  level were  identified  to morphospecies.   Density  (individuals  0.1 m‐2),  number  of  taxa  and  Simpson’s  diversity  index  (1‐λ`)  (Simpson  1949) were calculated to summarise different aspects of diversity.  These metrics were  the mean values calculated for the five individual samples at each site.        Figure 1: Location of 47 sites in the Tongariro National Park, central North Island, New Zealand,  sampled between February and April 2007.        Periphyton  biomass,  assessed  as  chlorophyll  a, was  used  as  a  surrogate  for  primary  productivity.    Morin  et  al.  (1999)  reviewed  the  relationship  between  chlorophyll a and productivity in streams and found a strong link between the two (r2 =  0.63).   Moreover, Tonkin and Death  (In prep.)  found a  strong  link between biomass  accumulation on artificial substrate and chlorophyll a on natural substrates (r2 = 0.74)  in these streams.   Periphyton biomass was estimated from measures of chlorophyll a  from five stones (mean area: 60 cm2) collected randomly from each site.  Stones were  kept  cool  in  the dark before being  frozen.   Chlorophyll a was  extracted using  90 %  Chapter 2: Interactive effects of productivity and disturbance on stream diversity      26    acetone at 5°C for 24 h in the dark.  Absorbances were read on a Varian Cary 50 conc  UV‐Visible  Spectrophotometer  (Varian  Australia  Pty  Ltd,  Mulgrave,  Australia)  and  converted  to pigment concentration  following Steinman and Lamberti  (1996).   Stone  surface area was estimated following Graham et al. (1988) and then halved to correct  for the proportion of the stone available for periphyton growth.      Physicochemical and substrate variables  Bed  stability/substrate  disturbance was  assessed  using  the  Pfankuch  stability  index  (Pfankuch  1975).    Only  the  bottom  component  of  the  index  (rock  angularity,  brightness,  packing,  percent  stable  materials,  scouring,  and  amount  of  clinging  vegetation) was  used,  as  this  is more  relevant  to  stream  invertebrate  communities  (Winterbourn and Collier 1987, Death and Winterbourn 1994).    Substrate  size  composition  was  assessed  using  the  ‘Wolman Walk’ method  where the beta axis of 100 stones was measured at approximately 1 m intervals across  a  zigzag  transect  at  45˚  to  the  stream  bank  (Wolman  1954).    Percentage  substrate  composition of Wentworth scale classes was converted to a single substrate size index  by summing midpoint values of size classes weighted by their proportion.  Bedrock was  assigned a nominal size of 400 mm for use in the calculations (Quinn and Hickey 1990).    Conductivity,  temperature  and  pH were measured  using  Eutech  instruments  ECScan  pocket meter.    Depth  and  velocity  were  recorded  with  a Marsh‐McBirney  flowmate  current meter  in  the  thalweg  of  each  stream  at  five  equidistant  intervals  along the study reach.   Flow type of each site was assessed visually as percentage of  backwater, pool, run or riffle over a 100 m reach.   Coarse particulate organic matter  (CPOM) was assessed visually as the percent bed cover of  leaf  litter within the 50 m  reach.    Fine  particulate  organic matter  (FPOM) was  not measured  as  POM  has  not  been shown  to be an  important determinate of  invertebrate community structure  in  New  Zealand  streams  (Scrimgeour  and Winterbourn  1989,  Death  and Winterbourn  1995).   Riparian vegetation percent composition  (native  forest, native scrub, planted  forest, pasture and bare ground) and % canopy cover over  the stream channel were  also assessed visually.    Chapter 2: Interactive effects of productivity and disturbance on stream diversity      27    In  order  to  test  for  differences  in  the  disturbance‐productivity‐diversity  relationship between open canopy and sites with canopy cover, sites were divided  in  half to create an open canopy and a canopy cover group differentiated at the median  value of around 30 % cover.   Specifically, open canopy sites (n = 24) were those with  less than 30 % overhead cover and sites with canopy cover (n = 23) those with greater  than or equal  to 30 % overhead cover, 17 of which had >50 % cover.   These groups  were kept for all analyses, as well as carrying out analyses on the complete data set.      Statistical analysis  To  explore  differences  in  community  structure  between  open  and  closed  canopy  streams we carried out analysis of  similarities  (ANOSIM)  (Clarke 1993) on  log  (x + 1)  data using Bray‐Curtis similarity in Primer v6 (Clarke and Gorley 2006).  We then used  similarity percentages (SIMPER) (Clarke 1993) to explore which taxa contributed to the  differences between canopy (Clarke and Gorley 2006).    In  order  to  test  for  differences  in  physicochemical  variables,  periphyton  and  diversity between open canopy and sites with canopy cover, we carried out analysis of  variance  (ANOVA)  using  Statistix  (Statistix  8 ©, Analytical  Software,  Tallahassee,  FL,  U.S.A.).    Differences  in  productivity  between  canopy  types  were  determined  with  analysis of covariance (ANCOVA) with disturbance and stream width as covariates.    If  required, data were  log (x + 1) transformed to adjust for normality.   Regressions and  ANCOVA between disturbance, productivity and diversity were carried out  in Statistix  (Statistix 8 ©, Analytical Software, Tallahassee, FL, U.S.A.).    The DEM predicts a unimodal relationship between diversity and productivity,  disturbance and the interaction between productivity and disturbance.  To test this we  fitted our data to the following polynomial:     S = b0 + b1P + b2C + b3PP 2+ b C4 2+b PxC 5   where P is the disturbance assessed as the Pfankuch index bottom component and C is  productivity assessed as chlorophyll a (µg cm‐2).   In order to explore the productivity‐ Chapter 2: Interactive effects of productivity and disturbance on stream diversity      28    disturbance‐diversity relationship further we developed the following models based on  regressions between the individual predictors and richness observed in the results:    Model 1  S = b0 + b1P + b2[ln(C)] + b3PP 2+b PxC 4 Model 2   S = b0 + b1P + b2[ln(C)] + b3PP 2    Model 1 has the  interaction term between disturbance and productivity  included and  model  2  has  the  interaction  excluded  in  order  to  explore  the  interactive  effects  of  isturbance.   productivity and d   Results  Physicochemical conditions  Conductivity ranged from 40 to 298 µS cm‐1 and was higher in open than closed canopy  streams (Table 1) but did not differ with the source of streams (i.e. runoff‐fed, spring‐ fed, or dammed) (F2, 45 = 0.92, P = 0.41).  Spot temperature ranged from 6.6 to 17.6˚C  at all sites and was slightly lower in spring‐fed streams (F2, 45 = 3.29, P = 0.047) but did  not differ with canopy cover (Table 1).  Mean velocity and depth ranged from 0.16 to  1.46 m s‐1 and 5.7 to 52.2 cm, respectively, and were greater in open canopy streams  (Table 1).  Overhead cover and stream width were negatively correlated (r = ‐0.45, P =  0.001), thus width was greater at open canopy sites (Table 1).      Taxonomic composition  The  presence  or  absence  of  canopy  cover was  a  strong  structuring  force  acting  on  community structure in these streams (ANOSIM R = 0.137, P = 0.001).  Ephemeroptera,  Plecoptera, Trichoptera and Chironomidae were  the dominant  taxa  in all study sites.   The mayfly Deleatidium spp.,  the stonefly Zelandoperla spp.,  the elmid beetle  larvae  Hydora spp. and chironomid Maoridiamesa spp. were abundant at both open canopy  and  sites with  canopy  cover  (>3.7 %  contribution  at  both  canopy  treatments).   No  Chapter 2: Interactive effects of productivity and disturbance on stream diversity      29    single  taxon  contributed more  than  5 %  to  the  difference  in  community  structure  between open  and  closed  canopy  streams  (SIMPER:  avg. dissimilarity  =  58.7).    Four  chironomids  were  amongst  the  top  five  contributors  to  differences  in  community  structure  between  open  and  closed  sites  (>3.5 %  contribution  to  differences).    The  filter  feeding mayfly  Coloburiscus  humeralis was  strongly  associated with  sites with  canopy cover (5.3 % contribution) as were Oligochaeta (5.1 % contribution), the mayfly  Austroclima  sepia  (3.2 %  contribution)  and  cased  caddisfly Beraeoptera  roria  (3.6 %  contribution).    The  net‐spinning  caddisfly Aoteapsyche  colonica  (6.4 %  contribution)  and  the  tipulid  Aphrophila  neozelandica  (3.3  %  contribution)  were  found  more  commonly at open canopy sites.      Table  1:  Mean  (±  1  S.E.)  physicochemical,  periphyton  and  invertebrate  community  characteristics for open and closed canopy streams collected from 47 streams in the Tongariro  National Park, New Zealand, between February and April 2007.   F and P values  for one‐way  ANOVA  testing  for  differences  between  open  and  canopy  streams  are  also  given.    SI  =  substrate size index, CPOM = coarse particulate organic matter.       Open canopy  Closed canopy  F1, 45 P  Width (m)  12.9 (1.95)  5.63 (1.03)  10.63  0.002  Depth (cm)  31.05 (1.85)  23.16 (2.42)  6.77  0.013  Velocity (m s‐1)  0.89 (0.04)  0.67 (0.08)  6.53  0.014  Conductivity (µS cm‐1)  131.58 (14.15)  93.17 (7.66)  5.56  0.023  Temperature (˚C)  10.71 (0.45)  10.87 (0.44)  0.06  0.8  pH  7.88 (0.12)  8.03 (0.06)  1.14  0.29  SI  152.59 (9.02)  134.98 (11.21)  1.51  0.23  CPOM (%)  2.46 (0.99)  12.48 (2.93)  10.88  0.002  Slope (m/10 m)  3.77 (0.35)  4.87 (0.5)  3.34  0.074  Pfankuch score  36.88 (2.12)  30.09 (2.42)  4.48  0.04  Chlorophyll a (µg cm‐2)  1.88 (0.3)  1.87 (0.23)  <0.01  0.98  No. of taxa 0.1 m‐2 14.78 (1.02)  18.02 (0.91)  5.6  0.022  No of individuals 0.1 m‐2 327.5 (59.52)  275.92 (32.15)  0.57  0.46  Simpson's (1 ‐ λ`)  0.74 (0.02)  0.81 (0.01)  9.52  0.004      Periphyton and stability  Chlorophyll a ranged from 0.03 to 5.02 µg cm‐2 and averaged 1.87 µg cm‐2.  There was  no difference  in chlorophyll a between open and closed sites (Table 1), and providing  Chapter 2: Interactive effects of productivity and disturbance on stream diversity      30    stream width as a covariate did not change this (ANCOVA: F1, 44 = 0.33, P = 0.57).  Sites  with canopy cover were more stable  (Pfankuch  index) but differences  in site stability  did not account for the lack of difference in periphyton biomass between open canopy  and sites with canopy cover (ANCOVA: F1, 44 = 0.14, P = 0.71).  Periphyton biomass was  not related to stability (Pfankuch index) for all sites combined (F1, 45 = 1.21, P = 0.28, r 2  = 0.026; Fig. 2) or for only sites with canopy cover (F1, 21 = 0.34, P = 0.57, r 2 = 0.02; Fig.  2), but declined with decreasing stability at open sites (F1, 22 = 4.46, P = 0.05, r 2 = 0.17;  Fig. 2).  In fact, multiple correlations show no single variable was related to periphyton  biomass with all sites  included.   CPOM was not related to stability at all sites (F1, 45 =  0.46, P = 0.5), open canopy sites (F1, 22 = 0.62, P = 0.44) or sites with canopy cover (F1,  21 = 2.73, P = 0.11).      Figure 2: Mean (± 1 S.E.) chlorophyll a as a function of the bottom component of the Pfankuch  stability  index  collected  from  47  streams  in  the  Tongariro  National  Park,  New  Zealand,  between February and April 2007.   Open circles are open canopy streams and closed circles  are sites with canopy cover.  Regression line represents only open canopy sites; see text for fit  statistics and equation.      Density and diversity  The mean total density of invertebrates in the benthos ranged from 3.8 individuals 0.1  m‐2 at  site 32  to 1205  individuals 0.1 m‐2 at  site 29.   Density did not differ between  Chapter 2: Interactive effects of productivity and disturbance on stream diversity      31    open canopy and sites with canopy cover (Table 1).   Density  increased logarithmically  with increasing periphyton biomass and declined at an increasing rate with increasing  disturbance levels (Table 2).  If open canopy sites were considered separately, density  also increased logarithmically with increasing periphyton biomass (Fig. 3; Table 2) and  peaked at low to intermediate levels of stability (Fig. 3; Table 2; AIC: ‐33.3) rather than  declining monotonically (F1, 22 = 12.92, P = 0.001, r 2 = 0.37; AIC: ‐32.4).  No relationship  with periphyton biomass or  stability was  evident  at  sites with  canopy  cover  (Fig.  3;  Table 2).  The number of individuals was not related to CPOM at all sites (F1, 45 = 0.04, P  = 0.84), open sites (F1, 22 = 0.01, P = 0.94) or sites with canopy cover (F1, 21 = 0.03, P =  0.87).      Table  2:  Results  of  regression  analysis  for mean  number  of  individuals  (density)  and mean  number of taxa as a function of a) bottom component of the Pfankuch stability  index and b)  chlorophyll a (µg cm‐2) collected from 47 streams in the Tongariro National Park, New Zealand,  between February and April 2007.  D.f. = degrees of freedom.      D.f.  F P r2 Equation  a) Pfankuch    Density  2, 44  6.21 0.004 0.23 y = 1.74 + 0.054x – 0.001x2       Open canopy 2, 21  9.43 0.001 0.47 y = 1.64 + 0.07x – 0.001x2       Closed canopy 1, 21  0.01 0.91 0.0006 Non‐significant  No. of taxa  2, 44  7.01 0.002 0.24 y = 9.90 + 0.64x – 0.01x2       Open canopy 2, 21  13.42 0.0002 0.56 y = ‐10.59 + 1.76x – 0.03x2       Closed canopy 1, 21  1.20 0.29 0.05 Non‐significant  b) Chlorophyll a   Density  1, 45  42.23 <0.0001 0.48 y = 2.22 + 0.28·ln(x)       Open canopy 1, 22  36.07 <0.0001 0.62 y = 2.24 + 0.31·ln(x)       Closed canopy 1, 21  2.99 0.099 0.13 Non‐significant  No. of taxa  1, 45  25.54 <0.0001 0.36 y = 15.48 + 2.97·ln(x)       Open canopy 1, 22  22.93 <0.0001 0.51 y = 14.41 + 2.79·ln(x)       Closed canopy 1, 21  4.22 0.053 0.17 Non‐significant      The mean  number  of  taxa  collected  per  sample  averaged  16.4  taxa  0.1 m‐2  ranging from 2.2 taxa 0.1 m‐2 to 27.8 taxa 0.1 m‐2.   Taxonomic richness was higher at  sites  with  canopy  cover  than  at  open  canopy  sites  (Table  1).    Richness  increased  logarithmically with  increasing periphyton biomass and peaked at  intermediate  levels  of  disturbance  (Table  2).   With  open  canopy  sites  considered  separately,  richness  Chapter 2: Interactive effects of productivity and disturbance on stream diversity      32    increased  logarithmically  with  increasing  periphyton  biomass  and  peaked  at  intermediate  levels of disturbance  (Fig. 4; Table 2).   No  relationship with periphyton  biomass or stability was found at sites with canopy cover (Fig. 4; Table 2).  The number  of taxa was not related to CPOM at all sites (F1, 45 = 3.79, P = 0.058), open sites (F1, 22 =  1.47, P = 0.24) or sites with canopy cover (F1, 21 = 0.49, P = 0.49).      Figure 3: Mean (± 1 S.E.) log number of animals as a function of (a, b) chlorophyll a and (c, d)  bottom  component  of  Pfankuch  stability  index  collected  from  47  streams  in  the  Tongariro  National Park, New Zealand, between February and April 2007.   (a, c) Open canopy sites and  (b, d) sites with canopy cover.  See Table 2 for regression results.      Chapter 2: Interactive effects of productivity and disturbance on stream diversity      33      Figure 4: Mean (± 1 S.E.) number of taxa as a function of (a, b) chlorophyll a and (c, d) bottom  component of Pfankuch  stability  index  collected  from 47  streams  in  the  Tongariro National  Park, New Zealand, between February and April 2007.  (a, c) Open canopy sites and (b, d) sites  with canopy cover.  See Table 2 for regression results.                Chapter 2: Interactive effects of productivity and disturbance on stream diversity      34    Simpson’s  diversity  index  was  higher  at  sites  with  canopy  cover  than  open  canopy  sites  (Table  1),  and  declined  with  decreasing  stability  when  all  sites  were  considered  together  (F1, 45  =  6.94,  P  =  0.01,  r 2  =  0.13;  y  =  0.866  –  0.00263x).    This  declining  trend  in Simpson’s  index with  stability was  significant at  sites with  canopy  cover  (F1, 45 = 13.15, P = 0.002, r 2 = 0.39; y = 0.9  ‐ 0.00291x) but not at open canopy  sites (F1, 45 = 0.14, P = 0.71, r 2 = 0.006).  Simpson’s index was not related to periphyton  biomass at all sites (F1, 45 = 0.05, P = 0.83, r 2 = 0.001), open canopy sites (F1, 45 = 0.0006,  P = 0.98,  r2 = 0.00003) or  sites with canopy cover  (F1, 45 = 0.28, P = 0.6,  r 2 = 0.013).   Simpson’s index was not related to CPOM at all sites (F1, 45 = 0.08, P = 0.78), open sites  (F1, 22 = 0.03, P = 0.86) or sites with canopy cover (F1, 21 = 2.88, P = 0.1).      Model testing  Fitting all sites to the DEM (S = b0 + b1P + b2C + b3PP 2+ b C4 2+b PxC) explained 42 % of  the variation  in the data (F = 5.84, P = 0.0004, r 5 5, 41  2 = 0.42; Fig. 5; Table 3), however  the only significant coefficient was the intercept (Table 3).   For only open canopy sites  the DEM explained 63 % of the variation in richness (F  = 6.07, P = 0.002, r5, 18 2 = 0.63;  Fig. 6; Table 3) with only the Pfankuch and Pfankuch2 coefficients significant.  The DEM  did not fit the data at sites with canopy cover (F  = 1.88, P = 0.15, r5, 17 2 = 0.36).    Model 1, with the interaction between productivity and disturbance included (S  =  b0  +  b1P  +  b2[ln(C)]  +  b3PP 2+b PxC),  explained  70 %  of  the  variation  in  the  open  canopy data (F  = 11.21, P < 0.0001, r 4 4, 19 2 = 0.70; Fig. 6; Table 3) and the same amount  of variation (42 %) as in the DEM for all sites (F  = 7.6, P = 0.0001, r4, 42 2 = 0.42; Fig. 5;  Table 3).   However, the  interaction between productivity and disturbance  in both the  open canopy sites and the all sites set was not significant.  There was no fit at the sites  with canopy cover (F  = 2.41, P = 0.09, r5, 17 2 = 0.35).    Model 2 with the interaction between productivity and disturbance excluded (S  = b0 + b1P + b2[ln(C)] + b3PP 2) decreased the overall fit slightly to 68 % of the variation in  richness  at  open  canopy  sites  but  the  three  coefficients  other  than  intercept were  significant  (F  = 14.36, P < 0.0001,  r3, 20 2 = 0.68;  Fig. 6;  Table 3).    The  fit once more  stayed the same as the previous two models with all data  included (F  = 10.26, P < 3, 43 Chapter 2: Interactive effects of productivity and disturbance on stream diversity      35    0.0001,  r2 = 0.42; Fig. 5; Table 3), and did not  fit at  sites with canopy cover  (F  =  1.69, P = 0.2, r 5, 17 2 = 0.21).      Table 3: Coefficients for the three models of taxonomic richness as a function of chlorophyll a  and  the  bottom  component  of  the  Pfankuch  stability  index  on  invertebrate  communities  collected from 47 streams in the Tongariro National Park, New Zealand, between February and  April 2007.  (a) DEM (S = b0 + b1P + b2C + b3P 2+ b4C 2+b5PxC), (b) model 1 – with productivity x  disturbance  interaction  (S  =  b0  +  b1P  +  b2[ln(C)]  +  b3P 2+b4PxC)  and  (c) model  2  – without  productivity x disturbance interaction (S = b0 + b1P + b2[ln(C)] + b3P 2) where P = Pfankuch and  C = chlorophyll a (µg cm‐2).      All sites  Open canopy sites    Model  Coefficient (SE)  t P Coefficient (SE) t  P  (a) DEM    b0 19.09 (6.99)  2.73 <0.01 ‐11.35 (14.06) ‐0.81  0.43 b1 ‐0.03 (0.39)  ‐0.09 0.93 1.50 (0.67) 2.24  0.04 b2 ‐1.20 (3.01)  ‐0.40 0.69 3.14 (4.69) 0.67  0.51 b3 ‐0.004 (0.005)  ‐0.67 0.51 ‐0.02 (0.008) ‐2.70  0.01 b4 ‐0.15 (0.38)  ‐0.40 0.69 ‐0.45 (0.43) ‐1.05  0.31 b5 0.10 (0.05)  1.87 0.07 ‐0.002 (0.09) ‐0.02  0.98 (b) Model 1 (interaction)    b0 15.19 (5.53)  2.75 <0.01 - 9.03 (7.50) - 1.20   0.24 b1 0.15 (0.37)  0.40 0.69 1.55 (0.46) 3.38   <0.01 b2 1.88 (1.28)  1.47 0.15 2.80 (1.11) 2.52  0.02 b3 ‐0.004 (0.005)  ‐0.76 0.45 - 0.02 (0.006 ) ‐3.48  <0.01 b4 0.01 (0.03)  0.46 0.65 - 0.02 (0.02) ‐1.11  0.28  (c) Model 2 (no interaction)    b0 14.6 (5.32)  2.74 <0.01 ‐6.56 (7.22) ‐0.91  0.37 b1 0.20 (0.35)  0.57 0.58 1.36 (0.43) 3.18  <0.01 b2 2.38 (0.66)  3.40 <0.001 1.79 (0.65) 2.77  0.01 b3 ‐0.004 (0.005)  ‐0.85 0.4 ‐0.02 (0.006) ‐3.28  <0.01         Chapter 2: Interactive effects of productivity and disturbance on stream diversity      36      Figure 5: Taxonomic richness as a function of chlorophyll a and the bottom component of the  Pfankuch stability index for open sites and sites with canopy cover collected from 47 streams  in the Tongariro National Park, New Zealand, between February and April 2007.  (a) DEM (S =  b0 + b1P + b2C + b3P 2+ b4C 2+b5PxC), (b) model 1 – with productivity x disturbance interaction (S  =  b0  +  b1P  +  b2[ln(C)]  +  b3P 2+b4PxC)  and  (c) model  2  – without  productivity  x  disturbance  interaction  (S = b0 + b1P + b2[ln(C)] + b3P 2) where P = Pfankuch and C = chlorophyll a.   See  Table 3 for coefficients and text for model fit.        Chapter 2: Interactive effects of productivity and disturbance on stream diversity      37        Figure 6: Taxonomic richness as a function of chlorophyll a and the bottom component of the  Pfankuch stability index for open canopy sites only collected from 47 streams in the Tongariro  National Park, New Zealand, between February and April 2007.  (a) DEM (S = b0 + b1P + b2C +  b3P 2+ b4C 2+b5PxC),  (b) model 1 – with productivity x disturbance  interaction  (S = b0 + b1P +  b2[ln(C)] + b3P 2+b4PxC) and (c) model 2 – without productivity x disturbance interaction (S = b0  + b1P + b2[ln(C)] + b3P 2) where P = Pfankuch and C = chlorophyll a.  See Table 3 for coefficients  and text for model fit.      Chapter 2: Interactive effects of productivity and disturbance on stream diversity      Discussion  38    Although the DEM explained some of the variation in benthic invertebrate richness (42  % with  all  sites  included),  the underlying  relationships did not  entirely match  those  predicted  by  the  DEM.    Like  Svensson  et  al.  (2007), we  found  diversity  peaked  at  intermediate  levels  of  disturbance,  measured  here  as  stream  bottom  stability  (Pfankuch  index),  irrespective of productivity  level; matching  the  trend predicted by  the  IDH  (Grime  1973b,  Connell  1978,  Sousa  1979).   Whilst  remaining  a  dominant  paradigm in disturbance ecology, there has been limited support for the IDH in streams  (but  see  Townsend et  al. 1997)  and  less  than 20 %  support  in  all ecological  studies  testing  this  theory  (Mackey  and  Currie  2001).    Similarly,  little  evidence  exists  supporting  the  DEM  in  streams,  although  Cardinale  et  al.  (2006)  found  support  in  primary  producer  communities  in  a  wide  variety  of  mid‐Atlantic,  North  American  streams.    The  only  effect  productivity  had  on  this  relationship  in  our  study was  in  determining the magnitude of the peak of the richness curve.       The role of canopy cover  The  relationship  between  richness,  stability  and  productivity  in  these  mountain  streams was  dictated  by  overhead  cover.    The  fit  to  the  DEM was  better  at  open  canopy sites than sites with canopy cover, and richness was related to productivity and  stability  only  at  open  canopy  sites.    In  contrast  to  the  patterns  observed  with  taxonomic  richness, Simpson’s diversity  index declined with decreasing  stability only  for streams with canopy cover.   Although overhead cover was related to stream size,  providing  stream  size  as  a  covariate  did  not  alter  the  relationships  found  between  open  sites  and  sites with  canopy  cover.    Previous  studies  have  found  that  factors  associated with  canopy  cover  can alter  the effects of  floods on  stream  invertebrate  communities  (likely  through  the  limitation of available  light)  (Robinson and Minshall  1986, Death 2002, Death and Zimmermann 2005, Fuller et al. 2008).   Most of  these  studies  also  found  that  floods  reduce periphyton  and  invertebrate diversity  at open  canopy  sites,  while  at  sites  with  canopy  cover,  periphyton  and  invertebrate  communities showed  little or no change (Death 2002, Death and Zimmermann 2005,  Chapter 2: Interactive effects of productivity and disturbance on stream diversity      39    Fuller et al. 2008).  In the absence of canopy, periphyton is the dominant food source  for invertebrates in stony streams whereas under canopy it is one of several potential  food sources.  Thus invertebrates would be expected to respond more strongly to algal  productivity gradients  in open canopy streams, rather than those with canopy cover.   The  high  mobility  of  lotic  invertebrates  allows  for  rapid  recolonisation  after  disturbance  but  is  dependent,  amongst  other  things,  on  resource  recovery  (e.g.  periphyton)  (Mackay  1992, Allan  1995, Death  and  Zimmermann  2005).   As  such,  in  streams with canopy cover where periphyton resources are  low, rate of recovery will  be  independent  of  algal  productivity.    Fuller  et  al.  (2008)  also  suggest  that  when  periphyton  recovery  is  not  limited  by  nutrients,  open  canopy  stream  invertebrate  communities are more resilient to disturbances.  The streams in this study are likely to  be  nutrient  limited  as  they  are  pristine  mountain  streams  so  we  would  expect  periphyton  recovery  to  be  slower  than  in  nutrient  enriched  streams  after  any  disturbance.    Although the density of overhead canopy in many New Zealand forests means  periphyton standing crops are often  low (Winterbourn and Fegley 1989, Winterbourn  1990,  Death  and  Zimmermann  2005),  there  was  no  effect  of  canopy  cover  on  productivity  in  the  streams  studied  here.    This may  be  due  to  recent  unrecorded  disturbance events removing periphyton biomass although this was not evident in the  time  leading  up  to  sampling  (pers.  obs.).    Alternatively,  it  could  be  a  result  of  differences  in  algal  quality  between  open  and  closed  canopy  streams  such  as  that  found with communities dominated by grazing resistant basal cells of the filamentous  green alga Stigeoclonium in the absence of light (Steinman et al. 1990, Barquin 2004).   Nonetheless, productivity declined with disturbance at open sites but not at sites with  canopy cover.   This may be due,  in part, to the  fact that the sites with canopy cover  were more stable than the open sites, potentially due to stabilisation of river banks by  riparian trees.  Fuller et al. (2008) suggest that canopy cover may lead to communities  more  resistant  to  pulsed  disturbances  as  long  as  organic matter  is  retained  during  floods or  that as much organic matter  is washed  in as  is washed out.   We  found no  relationship  between  stability  and  CPOM,  or  between  CPOM  and  invertebrate  community  diversity.    We  did  not  measure  FPOM,  which  is  generally  considered  Chapter 2: Interactive effects of productivity and disturbance on stream diversity      40    unimportant  in New Zealand streams  (Scrimgeour and Winterbourn 1989, Death and  Winterbourn  1995),  and  several  studies  have  found  that  POM  is  not  related  to  the  stability  of  both  open  and  closed  canopy  New  Zealand  streams  (Scrimgeour  and  Winterbourn  1989,  Death  and  Zimmermann  2005).    This  may  be  a  result  of  the  comparatively  low  level of  allocthonous material  entering or being  retained  in New  Zealand  streams  which  in  turn  causes  a  distinct  lack  of  obligate  shredders  (Winterbourn et al. 1981, Winterbourn 1997).    Lack of  trophic  specialisation  in New  Zealand stream  invertebrates generally, and the dependence on a trade‐off between  traits such as competitive and colonising ability (Chesson and Huntly 1997, Roxburgh et  al.  2004,  Cadotte  2007), makes  the  task  of  fitting models  to  these  communities  a  challenge.       Specific effects of disturbance and productivity  As noted earlier, diversity peaked at intermediate levels of disturbance as predicted by  the DEM.  The most widely accepted view of how disturbance affects diversity is that it  creates  new  niche  opportunities  by  removing  taxa  and  interrupting  biological  processes  such  as  competitive exclusion  (Connell 1978, Huston 1979, 1994, Cadotte  2007).   These opportunities allow  for  the greatest expression of  life history  traits at  intermediate  levels  of  disturbance.    Conversely,  in  systems  where  competitive  exclusion  is  less  prevalent  or  absent,  disturbance  may  simply  be  resetting  the  colonisation process by removing animals and their resource supply (Death 2002).  This  has been shown in streams with productivity setting the upper limit to richness rather  than leading to competitive exclusion, and in turn leading to a different relationship to  that  predicted  by  the  DEM  (Death  2002).    Evidence  from  our  study  suggests  that  disturbance  is  acting  along  the  lines  of  the  DEM/IDH  by  maximising  richness  at  intermediate levels of disturbance.  However, no taxa demonstrated strong trends for  either  end  of  the  disturbance  continuum;  what  appears  to  be  evident  is  a  large  proportion of taxa are common at intermediate disturbance sites but less so at either  stable  or  unstable  sites.    The  few  taxa  to  demonstrate  any  relationship  were  the  midges Maoridiamesa spp. and one species of Orthocladiinae, and the calocid caddisfly  Zelolessica  cheira  which  declined  with  disturbance  whereas  the  tipulid  Eriopterini  Chapter 2: Interactive effects of productivity and disturbance on stream diversity      41    increased.   Moreover,  the  taxa  exhibiting  trends  for  either  end  of  the  disturbance  continuum appear to be influenced by specific habitat preferences rather than due to  a competitive hierarchy.  For example, Zelolessica cheira favours bryophytes which are  associated with stable spring‐fed streams and Maoridiamesa  is often associated with  cold  stable  temperature  regimes  of  spring‐fed  streams  (Barquin  and  Death  2006).   Sites  with  much  higher  proportion  of  bryophyte  cover  can  exclude  other  non‐ specialised taxa such as the commonly widespread Deleatidium (Suren 1991).   Given that most stream insects are highly mobile organisms, we would expect a  unimodal relationship with disturbance to be  less  likely than  if we were dealing with  sessile  organisms  (Wootton  1998).    Previous  studies  have  shown  that  stream  invertebrate  diversity  declines  with  increasing  rates  of  disturbance  (Death  and  Winterbourn  1995,  Death  2002).    Due  to  the  nature  of  streams,  and  the  lack  of  evidence  to  suggest  competitive  exclusion  occurs  in  these  systems, we  feel Death’s  (Death  2002)  assertion  that  disturbance  acts  by  resetting  the  colonisation  process  holds more theoretical application in this situation.    The  DEM  predicts  diversity  will  be  maximised  at  intermediate  levels  of  productivity;  however,  in  the  present  study,  richness  and  density  both  increased  logarithmically  with  increasing  rates  of  productivity.    This  logarithmic  increase  in  diversity  suggests,  rather  than  productivity  controlling  the  rate  of  displacement  of  inferior competitors by superior competitors as predicted by the DEM, productivity  is  setting the upper limit to the potential richness of a community, as proposed by Death  (2002).    When  accounting  for  evenness,  diversity  (Simpson’s)  was  not  related  to  productivity.  Both unimodal (e.g. Grime 1973a, Huston 1979, Rosenzweig 1992, 1995,  Leibold  1999, Waide  et  al.  1999, Mittelbach et  al.  2001)  and  linear  (e.g. Currie  and  Paquin 1987, Currie 1991, Abrams 1995, Waide et al. 1999, Gaston 2000, Mittelbach et  al. 2001)  relationships have been  found between productivity and diversity  (Abrams  1995, Waide et al. 1999, Mittelbach et al. 2001).  One possible explanation for the non‐ uniform responses may be the scale at which a system is observed (Chase and Leibold  2002).   Additionally,  as with disturbance,  if  the  community  comprises highly mobile  taxa in an open environment such as streams, it is likely that competitive exclusion will  not be invoked at higher productivity and thus unimodal relationships will be unlikely.    Chapter 2: Interactive effects of productivity and disturbance on stream diversity      42    Interactions between productivity and disturbance  Both  the  DEM  and  our  model  had  no  interaction  term  between  productivity  and  stability.    Using  protozoan  communities  in  the  laboratory,  Scholes  et  al.  (2005)  concluded  similarly  that  diversity  responded  to  both  productivity  and  disturbance  without any interacting effects.  The DEM predicts that both the productivity‐richness  and disturbance‐richness relationships are unimodal, and that the  level of one  factor  interacts with the other to influence where the peak of that relationship forms.  Thus,  productivity and disturbance theoretically  interact to allow for the greatest variety of  life history traits to be expressed.   Perhaps, because of an absence of a competition‐ colonisation trade‐off in these streams, this interaction cannot occur and the greatest  diversity is simply at high productivities where resources are allowing more animals to  coexist.   Haddad et al. (2008) suggests  it  is not necessarily a competition colonisation  trade‐off but intrinsic growth rates of species that is more important for the response  to disturbance.    Productivity  can  help  to  explain  the  contrasting  responses  of  diversity  to  disturbance in ecological communities (Huston 1979, Kondoh 2001), but the model we  developed  here  simply  suggests  that,  rather  than  shifting  the  peak  in  diversity,  productivity  logarithmically  increases  the  magnitude  of  the  unimodal  disturbance‐ diversity curve.   Essentially  richness peaks at  intermediate  levels of disturbance with  productivity  determining  the  magnitude  of  that  peak.    Without  the  interaction  between  disturbance  and  productivity  we  cannot  unequivocally  suggest  that  disturbance  not  only  acts  directly  upon  invertebrates  but  on  their  food  supply,  as  suggested by Death (2002) and Death and Zimmerman (2005).  This lack of interaction  was  surprising  for  two  reasons.    Firstly,  theoretically  these  should  go  hand‐in‐hand  with disturbances creating niches and productivity controlling the rate of colonisation.   Secondly, although important, we believe the main effect of disturbance in streams is  through the removal of the food supply (periphyton) rather than the direct removal of  animals (Death 2002).    This  study  was  conducted  in  low  nutrient  streams  where  the  range  of  productivity  was  limited  and  therefore  the  results  only  represent  part  of  the  productivity gradient found  in nature.   Alpine streams are often characterised by  low  Chapter 2: Interactive effects of productivity and disturbance on stream diversity      43    productivity  and  instability  (Milner  and  Petts  1994, Gafner  and  Robinson  2007).    In  fact,  nutrient  enrichment  in  alpine  streams  can  alter  the  response  of  stream  invertebrate  communities  to  disturbance  (Robinson  and Minshall  1986, Gafner  and  Robinson 2007).   Thus  it  is  important to carefully consider the range of productivities  encompassed in any study (Scholes et al. 2005).  Perhaps the crux of the issue here is  that we only encountered the lower end of the scale on the productivity continuum for  assessing  the  DEM.    Consequently,  the  logarithmic  increase  in  richness  with  productivity we found may merely be the peak in a greater unimodal relationship and  an  interaction between productivity  and disturbance may well occur.    Furthermore,  the relationship between productivity and richness is highly variable at smaller spatial  scales (Mittelbach et al. 2001, Field et al. 2009).      Conclusions  In  summary,  we  found  little  evidence  to  support  the  DEM  within  the  range  of  conditions encountered, and  there was no  indication of competitive displacement at  higher  levels  of  productivity  as  predicted  by  the  DEM.    Although  productivity  and  disturbance were  shown  to be  important determinants of  richness, we  suggest  that  they act independently of each other and are not sufficient to explain patterns in these  mountain streams in isolation of other factors.   It appears more complex multivariate  models  than  those  such  as  the DEM  are  required  to  adequately explain diversity  in  these natural systems.  Although the DEM and our model were sufficient in predicting  diversity  in  the absence of canopy cover, where canopy was present  their success  in  predicting  diversity  was  extremely  limited.    Our model  suggests  richness  peaks  at  intermediate  levels disturbance  in autotrophic streams with productivity determining  e peak of  the  curve but no  statistical evidence of  an  interaction with disturbance  gulating diversity patterns.    th re   Chapter 2: Interactive effects of productivity and disturbance on stream diversity      References  44    Abrams,  P. A.  1995. Monotonic  or  unimodal  diversity  productivity  gradients  ‐ What  does competition theory predict. Ecology 76:2019‐2027.  Allan, J. D. 1995. Stream Ecology: Structure and Function of Running Waters. Chapman  and Hall, London.  Barquin, J. 2004. Spatial patterns of invertebrate communities in spring and runoff‐fed  streams.  PhD  thesis.  PhD  thesis, Massey  University,  Palmerston  North,  New  Zealand.  Barquin,  J., and R. G. Death. 2006. Spatial patterns of macroinvertebrate diversity  in  New Zealand springbrooks and rhithral streams. Journal of the North American  Benthological Society 25:768‐786.  Cadotte, M. W. 2007. Competition‐colonization  trade‐offs and disturbance effects at  multiple scales. Ecology 88:823‐829.  Cardinale, B. J., H. Hillebrand, and D. F. Charles. 2006. Geographic patterns of diversity  in  streams  are  predicted  by  a  multivariate  model  of  disturbance  and  productivity. Journal of Ecology 94:609‐618.  Chase,  J.  M.,  and  M.  A.  Leibold.  2002.  Spatial  scale  dictates  the  productivity‐ biodiversity relationship. Nature 416:427‐430.  Chesson, P., and N. Huntly. 1997. The roles of harsh and fluctuating conditions  in the  dynamics of ecological communities. The American Naturalist 150:519‐553.  Clarke,  K.  R.  1993.  Nonparametric  multivariate  analyses  of  changes  in  community  structure. Australian Journal of Ecology 18:117‐143.  Clarke,  K.  R.,  and  R. N. Gorley.  2006.  PRIMER  v6: User Manual/Tutorial.  PRIMER‐E,  Plymouth.  Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs. Science 199:1302‐ 1310.  Currie, D. J. 1991. Energy and large‐scale patterns of animal‐species and plant‐species  richness. American Naturalist 137:27‐49.  Currie,  D.  J.,  and  V.  Paquin.  1987.  Large‐scale  biogeographical  patterns  of  species  richness of trees. Nature 329:326‐327.  Chapter 2: Interactive effects of productivity and disturbance on stream diversity      45    Death, R. G. 2002. Predicting invertebrate diversity from disturbance regimes in forest  streams. Oikos 97:18‐30.  Death, R. G. 2008. Effects of floods on aquatic invertebrate communities.in J. Lancaster  and R. A. Briers, editors. Insects: Challenges to Populations. CAB International,  UK, pp 103‐121.  Death,  R.  G.  2010.  Disturbance  and  riverine  benthic  communities:  What  has  it  contributed  to  general  ecological  theory?  River  Research  and  Applications  26:15‐25.  Death,  R. G.,  and M.  J. Winterbourn.  1994.  Environmental  stability  and  community  persistence:  a  multivariate  perspective.  Journal  of  the  North  American  Benthological Society 13:125‐139.  Death,  R.  G.,  and  M.  J.  Winterbourn.  1995.  Diversity  patterns  in  stream  benthic  invertebrate communities:  the  influence of habitat  stability. Ecology 76:1446‐ 1460.  Death,  R.  G.,  and  E. M.  Zimmermann.  2005.  Interaction  between  disturbance  and  primary  productivity  in  determining  stream  invertebrate  diversity.  Oikos  111:392‐402.  Field, R., B. A. Hawkins, H. V. Cornell, D. J. Currie, J. A. F. Diniz‐Filho, J. F. Guegan, D. M.  Kaufman,  J. T. Kerr, G. G. Mittelbach, T. Oberdorff, E. M. O'Brien, and  J. R. G.  Turner. 2009. Spatial species‐richness gradients across scales: a meta‐analysis.  Journal of Biogeography 36:132‐147.  Fuller, R. L., C. LaFave, M. Anastasi, J. Molina, H. Salcedo, and S. Ward. 2008. The role  of  canopy  cover  on  the  recovery  of  periphyton  and  macroinvertebrate  communities after a month‐long flood. Hydrobiologia 598:47‐57.  Gafner, K., and C. T. Robinson. 2007. Nutrient enrichment influences the responses of  stream  macroinvertebrates  to  disturbance.  Journal  of  the  North  American  Benthological Society 26:92‐102.  Gaston, K. J. 2000. Global patterns in biodiversity. Nature 405:220‐227.  Graham, A. A., D. J. McCaughan, and F. S. McKee. 1988. Measurement of surface area  of stones. Hydrobiologia 157:85‐87.  Chapter 2: Interactive effects of productivity and disturbance on stream diversity      46    Grime,  J. P. 1973a. Competitive exclusion  in herbaceous vegetation. Nature 242:344‐ 347.  Grime,  J.  P.  1973b.  Control  of  species  density  in  herbaceous  vegation.  Journal  of  Environmental Management 1:151‐167.  Gross, K., and B. J. Cardinale. 2007. Does species richness drive community production  or  vice  versa?  Reconciling  historical  and  contemporary  paradigms  in  competitive communities. American Naturalist 170:207‐220.  Haddad, N. M., M. Holyoak, T. M. Mata, K. F. Davies, B. A. Melbourne, and K. Preston.  2008.  Species'  traits  predict  the  effects  of  disturbance  and  productivity  on  diversity. Ecology Letters 11:348‐356.  Hastings,  A.  1980.  Disturbance,  coexistence,  history,  and  competition  for  space.  Theoretical Population Biology 18:363‐373.  Hubbell,  S.  P.  2001.  The  unified  neutral  theory  of  biodiversity  and  biogeography.  Princeton University Press.  Huston, M. 1979. A general hypothesis of  species diversity. The American Naturalist  113:81‐100.  Huston,  M.  1994.  Biological  Diversity:  The  Coexistence  of  Species  on  Changing  Landscapes. Cambridge University Press, Cambridge, UK.  Kondoh, M.  2001. Unifying  the  relationships  of  species  richness  to  productivity  and  disturbance.  Proceedings  of  the  Royal  Society  of  London  Series  B‐Biological  Sciences 268:269‐271.  Lake, P. S. 2000. Disturbance, patchiness, and diversity in streams. Journal of the North  American Benthological Society 19:573‐592.  Leibold,  M.  A.  1999.  Biodiversity  and  nutrient  enrichment  in  pond  plankton  communities. Evolutionary Ecology Research 1:73‐95.  Mackay, R. J. 1992. Colonization by Lotic Macroinvertebrates  ‐ a Review of Processes  and Patterns. Canadian Journal of Fisheries and Aquatic Sciences 49:617‐628.  Mackey,  R.  L.,  and  D.  J.  Currie.  2000.  A  re‐examination  of  the  expected  effects  of  disturbance on diversity. Oikos 88:483‐493.  Mackey,  R.  L.,  and  D.  J.  Currie.  2001.  The  diversity‐disturbance  relationship:  Is  it  generally strong and peaked? Ecology 82:3479‐3492.  Chapter 2: Interactive effects of productivity and disturbance on stream diversity      47    Milner,  A. M.,  and  G.  E.  Petts.  1994.  Glacial  rivers  ‐  physical  habitat  and  ecology.  Freshwater Biology 32:295‐307.  Mittelbach, G. G., C. F. Steiner, S. M. Scheiner, K. L. Gross, H. L. Reynolds, R. B. Waide,  M. R. Willig, S. I. Dodson, and L. Gough. 2001. What is the observed relationship  between species richness and productivity? Ecology 82:2381‐2396.  Morin, A., W. Lamourex, and  J. Busnarda. 1999. Empirical models predicting primary  productivity  from chlorophyll a and water temperature  for stream periphyton  and  lake  and  ocean  phytoplankton.  Journal  of  the  North  American  Benthological Society 18:299‐307.  Palmer,  M.  W.  1994.  Variation  in  species  richness  ‐  towards  a  unification  of  hypotheses. Folia Geobotanica & Phytotaxonomica 29:511‐530.  Pfankuch, D. J. 1975. Stream Reach  Inventory and Channel Stability Evaluation. USDA  Forest Service, Region 1, Missoula, Montana.  Pickett, S. T. A., and P. S. White. 1985. The ecology of natural disturbance and patch  dynamics. Academic Press.  Quinn, J. M., and C. W. Hickey. 1990. Magnitude and effects of substrate particle size,  recent  flooding,  and  catchment  development  on  benthic  invertebrates  in  88  New Zealand rivers. New Zealand  Journal of Marine and Freshwater Research  24:411‐427.  Reice, S. R. 1985. Experimental disturbance and the maintenance of species‐diversity in  a stream community. Oecologia 67:90‐97.  Resh, V. H., A. V. Brown, A. P. Covich, M. E. Gurtz, H. W. Li, G. W. Minshall, S. R. Reice,  A. L. Sheldon, J. B. Wallace, and R. C. Wissmar. 1988. The role of disturbance in  stream  ecology.  Journal  of  the North  American  Benthological  Society  7:433‐ 455.  Robinson, C. T., and G. W. Minshall. 1986. Effects of disturbance frequency on stream  benthic community structure in relation to canopy cover and season. Journal of  the North American Benthological Society 5:237‐248.  Rosenzweig, M. L. 1992. Species‐diversity gradients ‐ we know more and less than we  thought. Journal of Mammalogy 73:715‐730.  Chapter 2: Interactive effects of productivity and disturbance on stream diversity      48    Rosenzweig, M.  L.  1995.  Species Diversity  in  Space  and  Time.  Cambridge University  Press, Cambridge, UK.  Rosenzweig, M. L., and Z. Abramsky. 1993. How are Diversity and Productivity Related?  Pages  52‐65  in  R.  E.  Ricklefs  and  D.  Schluter,  editors.  Species  Diversity  in  Biological Communities. University of Chicago Press, Chicago, Illinois, USA.  Roxburgh,  S.  H.,  K.  Shea,  and  J.  B.  Wilson.  2004.  The  intermediate  disturbance  hypothesis:  Patch  dynamics  and mechanisms  of  species  coexistence.  Ecology  85:359‐371.  Scholes, L., P. H. Warren, and A. P. Beckerman. 2005. The combined effects of energy  and  disturbance  on  species  richness  in  protist  microcosms.  Ecology  Letters  8:730‐738.  Scrimgeour,  G.  J.,  and M.  J. Winterbourn.  1989.  Effects  of  floods  on  epilithon  and  benthic  macroinvertebrate  populations  in  an  unstable  New‐Zealand  river.  Hydrobiologia 171:33‐44.  Simpson, E. H. 1949. Measurement of diversity. Nature 163:688.  Sousa, W. P. 1979. Disturbance in marine intertidal boulder fields: the nonequilibrium  maintenance of species diversity. Ecology 60:1225‐1239.  Sousa, W. P. 1984. The role of disturbance  in natural communities. Annual Review of  Ecology and Systematics 15:353‐391.  Steinman, A. D.,  and G. A.  Lamberti.  1996. Biomass  and  pigments  of  benthic  algae.  Pages 295‐314  in F. R. Hauer and G. A. Lamberti, editors. Methods  in Stream  Ecology. Academic Press, San Diego, CA.  Steinman, A. D., P.  J. Mulholland, A. V. Palumbo, T. F. Flum,  J. W. Elwood, and D. L.  Deangelis.  1990.  Resistance  of  lotic  ecosystems  to  a  light  elimination  disturbance ‐ a laboratory stream study. Oikos 58:80‐90.  Sugden, H., M.  Lenz, M. Molis, M. Wahl, and  J. C. Thomason. 2008. The  interaction  between  nutrient  availability  and  disturbance  frequency  on  the  diversity  of  benthic marine  communities  on  the  north‐east  coast  of  England.  Journal  of  Animal Ecology 77:24‐31.  Suren,  A.  M.  1991.  Bryophytes  as  invertebrate  habitat  in  2  New  Zealand  alpine  streams. Freshwater Biology 26:399‐418.  Chapter 2: Interactive effects of productivity and disturbance on stream diversity      49    Svensson,  J. R., M. Lindegarth, M. Siccha, M. Lenz, M. Molis, M. Wahl, and H. Pavia.  2007. Maximum  species  richness  at  intermediate  frequencies of disturbance:  Consistency among levels of productivity. Ecology 88:830‐838.  Tilman, D. 1994. Competition and biodiversity in spatially structured habitats. Ecology  75:2‐16.  Tonkin,  J.  D.,  and  R.  G.  Death.  In  prep.  Consistent  effects  of  productivity  and  disturbance on diversity between landscapes.  Towns,  D.  R.,  and  W.  L.  Peters.  1996.  Leptophlebiidae  (Insecta:  Ephemeroptera).  Manaaki Whenua Press, Lincoln, New Zealand.  Townsend, C. R., M. R. Scarsbrook, and S. Doledec. 1997. The intermediate disturbance  hypothesis, refugia, and biodiversity  in streams. Limnology and Oceanography  42:938‐949.  Vinson, M. R., and C. P. Hawkins. 1998. Biodiversity of stream insects: variation at local,  basin, and regional scales. Annual Review of Entomology 43:271‐293.  Waide, R. B., M. R. Willig, C.  F.  Steiner, G. Mittelbach,  L. Gough,  S.  I. Dodson, G. P.  Juday,  and  R.  Parmenter.  1999.  The  relationship  between  productivity  and  species richness. Annual Review of Ecology and Systematics 30:257‐300.  Winterbourn, M.  J. 1990.  Interactions among nutrients, algae and  invertebrates  in a  New Zealand mountain stream. Freshwater Biology 23:463‐474.  Winterbourn, M.  J.  1997.  New  Zealand  mountain  stream  communities:  stable  yet  disturbed?  Pages  31‐54  in  B.  Streit,  T.  Stadler,  and  C.  M.  Lively,  editors.  Evolutionary  Ecology  of  Freshwater  Animals.  Birkhauser  Verlag,  Basel,  Switzerland.  Winterbourn, M. J., and K. J. Collier. 1987. Distribution of benthic invertebrates in acid,  brown  water  streams  in  the  South  Island  of  New  Zealand.  Hydrobiologia  153:277‐286.  Winterbourn, M. J., and A. Fegley. 1989. Effects of nutrient enrichment and grazing on  periphyton assemblages in some spring‐fed South Island streams. New Zealand  Journal of Natural Sciences 16:57‐65.  Winterbourn, M.  J., K.  L. D. Gregson, and C. H. Dolphin. 2000. Guide  to  the aquatic  insects of New Zealand. Entomological Society of New Zealand, Auckland.  Chapter 2: Interactive effects of productivity and disturbance on stream diversity      50    Winterbourn,  M.  J.,  J.  S.  Rounick,  and  B.  Cowie.  1981.  Are  New  Zealand  stream  ecosystems  really  different? New  Zealand  Journal  of Marine  and  Freshwater  Research 15:321‐328.  Wolman, M. J. 1954. A method of sampling coarse river bed material. Transactions of  the American Geophysical Union 35:951‐956.  Wootton,  J.  T.  1998.  Effects  of  disturbance  on  species  diversity:  A  multitrophic  perspective. American Naturalist 152:803‐825.      Chapter 3  Consistent effects of productivity and  disturbance on diversity between landscapes                         Chapter 3: Productivity, disturbance and diversity between landscapes      Abstract  52    Productivity  and disturbance have  a  strong  role  in determining diversity patterns  in  nature  yet  whether  they  operate  individually  or  interact  to  determine  diversity  is  unclear.  Moreover, what effect land‐use change has on this relationship has not been  assessed.    We  tested  whether  land‐use  influenced  the  relationship  between  productivity,  disturbance  and  diversity  using  data  from  multiple  samplings  of  16  streams  in  two  contrasting  regions  of  the  North  Island  of  New  Zealand.    As  the  Dynamic Equilibrium Model (DEM) has received inconsistent support in all ecosystems  and  little  favourable  applications  in  lotic  systems,  we  applied  this  along  with  two  previously  developed  for  stream  communities.    Although  the  community  structure  differed between the two regions, the response of taxonomic richness to productivity  and  disturbance  was  consistent.    That  is,  richness  responded  log‐linearly  to  productivity  and  declined monotonically with  disturbance.   However,  there was  no  evidence of an interactive effect of productivity and disturbance.  When accounting for  density  (rarefaction)  the  results  were  inconsistent,  exhibiting  no  relationship  with  productivity  but  declining  with  disturbance.    Our  results  suggest  the  Death  productivity‐disturbance‐diversity  is the most applicable model  in these communities  where disturbance  simply  removes  taxa and productivity  controls  the upper  limit  to  richness.      Chapter 3: Productivity, disturbance and diversity between landscapes      Introduction  53    The explanation of diversity patterns has been a  focus  for ecologists  for a  long  time  and continues to be a major challenge due to the multitude of explanatory factors (e.g.  Huston  1994, Hubbell 2001, Ricklefs  2004).   Consequently,  rather  than  single  factor  models,  ecologists  have  developed  ever more  complex models  to  explain  diversity  patterns.  Mackey and Currie (2001), in their review, highlight the fact that there is an  ever  lengthening  list  of  factors  that  can  influence  diversity  in  natural  systems,  but  there  has  been  little  progress  in  identifying  what  factors  explain  the  majority  of  variation  in diversity observed  in nature.   Emphasis has been on developing diversity  models based on abiotic controls, particularly disturbance interrupting processes such  as  competitive  exclusion.    Of  the  many  factors  that  have  been  hypothesised  to  determine  diversity  patterns  in  natural  environments,  disturbance  (Connell  1978,  Wootton  1998,  Lake  2000)  and  productivity  (Currie  1991,  Waide  et  al.  1999,  Mittelbach  et  al.  2001)  are  those most  often  invoked.   Often,  the  appeal  of  these  hypotheses  is  their simplicity; however,  this can also be  their weakness as  they may  only  apply  to  a  single  trophic  level  (Wootton  1998,  Mackey  and  Currie  2001).   Nevertheless,  although  these models  appear  simplistic  at  a  glance  they may  in  fact  account for multiple underlying mechanisms of coexistence (Roxburgh et al. 2004).    Two  of  the  most  longstanding  models  of  diversity  are  the  Intermediate  Disturbance  Hypothesis  (IDH)  (Connell  1978)  and  the  Dynamic  Equilibrium  Model  (DEM) (Huston 1979, 1994).  The IDH predicts diversity to peak at intermediate levels  of disturbance  and  the DEM,  and Kondoh’s extension of  this model  (Kondoh 2001),  builds  on  the  IDH  to  predict  a  peak  at  intermediate  levels  of  productivity  and  an  interaction  between  disturbance  and  productivity.    Several  empirical  studies  have  found  an  interaction  between  productivity  and  disturbance  to  be  important  in  controlling diversity (e.g. Wilson and Tilman 2002, Kneitel and Chase 2004, Death and  Zimmermann  2005,  Scholes  et  al.  2005).   However,  at  the  time  of  their  publication  Mackey and Currie (2000) found little evidence to suggest that disturbance should play  more than a minor role in determining diversity patterns in nature.    Chapter 3: Productivity, disturbance and diversity between landscapes      54    There has been limited support for any of the major models of diversity in lotic  systems (but see Townsend et al. 1997).  However, disturbance both alone (Death and  Winterbourn  1995,  Lake  2000),  and  in  combination with  productivity  (Death  2002,  Cardinale et al. 2005, Death and Zimmermann 2005, Cardinale et al. 2006), has been  found  to have a  strong  influence on diversity of  stream benthic  communities and  is  clearly one of the most  important structuring  forces of stream communities  (Resh et  al. 1988, Lake 2000, Death 2008).  Yet, there is little consensus on quite how diversity  responds to productivity and disturbance.  Streams provide a unique testing ground for  these  models  because  community  dynamics  are  dominated  by  immigration  and  emigration  rather  than  population  growth  for which most  theoretical  and  empirical  work  in  this  field has been applied.   Many of  the assumptions of  the major diversity  models are not met  in  lotic communities which are dominated by highly mobile taxa  (Hildrew and Giller 1994, Allan 1995) rather than sedentary organisms such as plants  and corals for which many of these models have been developed.    Death (2002) proposed a model that,  like the DEM, has diversity predicted by  the interaction between productivity and disturbance.  He found substrate disturbance  had  a  considerably  weaker  effect  on  macroinvertebrate  communities  in  forested  streams where light was limiting primary productivity.  He proposed that the principal  effect  of  disturbance  on  diversity  in  New  Zealand  streams  was  to  limit  primary  productivity.  This implies that productivity sets an upper limit to invertebrate diversity  in  streams  and disturbance  resets  the  colonisation process  of  stream  invertebrates.   This model predicts that diversity will increase log‐linearly with increasing productivity  and decline  linearly with  increasing disturbance.   Tonkin et al.  (In prep.) built on  this  model to suggest a quadratic decline in diversity with increasing disturbance and found  a stronger dependence on productivity and disturbance in open, as opposed to closed  canopy streams.    Although  there  is  evidence  of  an  interactive  effect  of  productivity  and  disturbance  on  diversity  in  numerous  ecological  systems  (Wilson  and  Tilman  2002,  Kneitel and Chase 2004, Scholes et al. 2005), the way in which this interaction operates  to affect diversity is not clear.  Cardinale et al. (2006) found evidence of this interaction  when applying the DEM to a wide range of North American streams, yet when testing  Chapter 3: Productivity, disturbance and diversity between landscapes      55    for this interaction on pristine New Zealand mountain streams, Tonkin et al. (In prep.),  found  no  such  interactive  effect  of  productivity  and  disturbance  on  diversity.   Essentially,  whether  the  effects  of  productivity  and  disturbance  on  diversity  are  additive  or  multiplicative  is  still  unclear.    We  would  expect  if  productivity  was  controlling the rate of recolonisation after disturbance that there should be significant  interaction terms.    Habitats worldwide  are under  threat  from  increasing  land use  intensification  (Tilman 1999b, Allan 2004, Foley et al. 2005).  Changes in land use such as agricultural  intensification  can  severely  alter  the  composition  and  biodiversity  of  communities  living within these landscapes (Harding et al. 1998, Allan 2004, Tscharntke et al. 2005).   These changes can lead to a subsidy‐stress response of communities where the initial  response may  be  positive  but  becomes  negative  at  greater  rates  of  environmental  change (Quinn 2000, Niyogi et al. 2007, Death and Collier 2010).  Not only will changes  in  land  use  alter  the  composition  of  communities,  but  in  streams,  alteration  of  catchment land use can lead to drastic changes in the flood regime (Rowe et al. 1997,  Walsh  et  al.  2005).    Pastoral  development  appears  to  influence  lotic  communities  through a  suite of mechanisms  such as  changes  to  sediment,  temperature and  light  regimes, channel morphology, hydrology, and the food base (Quinn 2000, Allan 2004).   However,  research on  the  influence of  agricultural  intensification on  flood‐effects  is  limited  (Death  2008),  but  Collier  and  Quinn  (2003)  found  that  pulse  disturbance,  through a major catchment‐scale  flood, can have differential effects on  forested and  pasture streams.   They suggest  that  interaction between press disturbance  (land‐use  change)  and  pulse  disturbance  (floods)  (Lake  2000)  can  have  strong  effects  on  the  post‐flood recovery of lotic communities.    The  effects  of  catchment  land  use  on  the  response  of  stream  invertebrate  communities to flow disturbance are not well documented.  As far as we know, no one  has  investigated whether varying composition of communities as a result of different  land  use  leads  to  differential  responses  of  diversity  patterns  to  productivity  and  disturbance.  Here we investigate whether the effects of productivity and disturbance  on diversity differ between a pristine and non‐pristine agricultural region by sampling  16 streams on multiple occasions  in the North Island of New Zealand.    In order to do  Chapter 3: Productivity, disturbance and diversity between landscapes      56    this we  fit these data to three models of productivity‐disturbance‐diversity: the DEM  (Huston  1979,  1994),  Death’s  (2002)  model  of  diversity  specifically  developed  for  stream  communities,  and Tonkin et  al.’s  (In prep.)  version of  this model.   Based on  Tonkin et al. (In prep.) and Death (2002) we predict diversity will  increase  log‐linearly  with productivity and decline  (linear or quadratic) with disturbance at pristine  sites.   We predict that non‐pristine sites will be affected by disturbance more severely than  pristine sites and diversity will decline at the higher levels productivity.  We expect the  change  in  land‐use  will  lead  to  clearer  interactions  between  productivity  and  disturbance due  to  the more extreme effects of  floods on  these  streams and higher  d with land‐use change to agriculture.   productivity associate   Methods  Study sites  The Tongariro National Park,  in the North  Island central plateau of New Zealand,  is a  world heritage site dominated by a central volcanic massif.  Rainfall ranges from 1,100‐ 3,500 mm  yr‐1  and  vegetation  varies  within  the  park  from mixed  beech‐podocarp  forest to native tussock and scrub.   Pastoral  farming and wine production dominates  the land‐use of Hawke’s Bay, in the East of the North Island, which is characterised by  a  warm  and  dry  climate  with  a  mean  annual  rainfall  of  783  mm  in  the  central  Ruataniwha Plains.   Eight streams were sampled  in each of  these  two  regions which  were  selected  to  represent  a  pristine  (Tongariro)  and  non‐pristine  (Hawke’s  Bay)  region.  Study sites were selected to represent high and low productivity in stable and  unstable  streams  in  each  region.  Classification  of  sites  as  high  or  low  productivity/stability was based on visual estimates of algal  cover, overhead  canopy  cover,  conductivity and  the Pfankuch bottom  component  (Pfankuch 1975).   Figure 1  shows the location of the 16 sites in the North Island of New Zealand.      Sampling protocol  Sampling took place on four occasions between February 2008 and August 2009.  Two  samplings took place in the austral summer and two in the austral winter in February  Chapter 3: Productivity, disturbance and diversity between landscapes      57    2008,  September  2008,  February  2009  and August  2009.   Macroinvertebrates were  sampled by taking five 0.1 m2 Surber samples (250 µm mesh) from random locations in  riffles throughout ~50 m reaches at each site.   Samples were stored  in 10%  formalin  and  later  identified  in  the  laboratory  to  the  lowest  possible  taxonomic  level  using  available keys (e.g. Towns and Peters 1996, Winterbourn et al. 2000).        Figure 1: Map showing location of the 16 study sites in two regions of the North Island of New  Zealand  collected  on  four  occasions  between  February  2008  and  July  2009.    a.  Tongariro  National  Park  (pristine)  and  b.  Hawke’s  Bay  (non‐pristine).    The  key  shows  the  difference  between stable and unstable and low and high productivity sites based on a priori selection.      Periphyton  To assess primary productivity we used a measure of biomass accumulation on  tiles  placed on the substrate.  One month prior to each of the four sampling occasions one  set of six unglazed terracotta tiles were deployed at each site.  Six 45 mm x 45 mm tiles  were attached to one 500 mm x 500 mm  interlocking rubber safety mat at each site.   The six  tiles were attached at evenly spaced  intervals on  the mats by drilling 10 mm  holes  in tiles and mats and attaching via 10 mm dowell.   Mats were secured  in place  Chapter 3: Productivity, disturbance and diversity between landscapes      58    flush with  the streambed with a metal  fencing pole and  long  tent pegs.   These mats  were left in place for one month prior to sampling.    As  a  second  estimate  of  primary  productivity we  also measured  biomass  on  natural  substrates at each  site by extracting chlorophyll a.   Periphyton chlorophyll a  and primary productivity  in  streams are highly  correlated; Morin et al.  (1999),  in an  extensive  review,  found a strong  log‐linear  relationship  (r2 = 0.63) between  the  two.   Five stones (α axis < 60 mm) were randomly collected from each riffle for extraction of  chlorophyll a.    Stones and tiles were kept cool and dark on ice in the field before being stored  at ‐20˚C.  Photosynthetic pigments were extracted from stones or tiles by submerging  in known volumes of 90% acetone for 24 hours at 5˚C.   Absorbances at 750, 665 and  664  nm were  read  on  a Varian Cary  50  conc UV‐Visible  Spectrophotometer  (Varian  Australia Pty Ltd, Mulgrave, Australia) and converted to chlorophyll a and phaeophytin  pigment  concentration  using  Steinman  and  Lamberti  (1996).    These  were  then  corrected for stone surface area and using Graham et al. (1988) and halved to account  for periphyton being present only on stone surfaces.      Measurement of disturbance  Bed  stability/substrate  disturbance was  assessed  using  the  Pfankuch  stability  index  (Pfankuch  1975).    Only  the  bottom  component  of  the  index  (which  evaluates  rock  angularity,  brightness,  packing,  percent  stable  materials,  scouring,  and  amount  of  clinging  vegetation)  was  used,  as  this  is  more  relevant  to  stream  invertebrate  communities  (Winterbourn  and  Collier  1987,  Death  and Winterbourn  1994).    We  supplemented this with direct measurement of stone movement at the streams.    Fifteen stones were used at each site, five from each of three size classes were  placed  in  sets of  three  (one of  each of  the  three  size  class)  at  equidistant  intervals  along the study reach (up to 100 m).  Stones were selected from the D50, D70 and D90  of individual streams.  Stones at each stream were marked with fluorescent spray paint  before  being  placed  in  the  bed.    These  tracer  particles were  placed  in  the  bed  as  closely mimicking the surrounding stones as possible without unnecessary disturbance  of the bed (i.e. embedded or sitting loosely).  Rather than measure distances moved by  Chapter 3: Productivity, disturbance and diversity between landscapes      59    individual  tracers  as  previous  studies  have  (e.g.  Downes  et  al.  1998,  Death  and  Zimmermann 2005), stones were simply recorded as moved or not, similar to that of  Townsend et al. (1997).  An index of bed movement is simply the percentage of overall  stone movement  for  each  stream.    Stone movement was measured  for  the month  leading up to biological sampling.      Other environmental measures  Substrate size composition was assessed using the ‘Wolman Walk’ method where the  β  axis  of  100  stones was measured  at  approximately  1 m  intervals  across  a  zigzag  transect at 45˚ to the stream bank (Wolman 1954).  Percentage substrate composition  of Wentworth scale classes was converted to a single substrate size index by summing  midpoint values of size classes weighted by their proportion.  Bedrock was assigned a  nominal size of 400 mm (Quinn and Hickey 1990).  Conductivity, temperature and pH  were measured using ECScan pocket meters.  Depth and velocity were recorded with a  Marsh‐McBirney flowmate current meter in the thalweg of each stream at five points  at equidistant  intervals  along  the  study  reach.   Flow  type of each  site was  assessed  visually  as  percentage  of  still,  backwater,  pool,  run  or  riffle  over  a  100  m  reach.   Percentage  of  leaf  litter,  riparian  vegetation  and  canopy  cover  were  also  visually assessed.     Community metrics  The  number  of  individuals  and  two  simple measures  of  diversity were  used  in  this  study: the number of taxa (richness) and rarefied species richness (ESN).   Rarefaction  accounts for the likelihood of capturing rare taxa with increased number of individuals  collected (Hurlbert 1971).  This is achieved by standardising the sites by predicting the  number of taxa expected for a fixed number of  individuals.    In order to get sufficient  numbers  for  the  index to be accurate  the  five replicates at each site were pooled  to  give one value  for each  site.   Five  site  samplings were  removed because of  the  low  number of collected animals and rarefied richness was calculated for 224 individuals.      Chapter 3: Productivity, disturbance and diversity between landscapes      60    Statistical analysis  The effects of productivity, stability and region based on a priori group selection were  assessed  using  three‐way  crossed  Analysis  of  Variance  (ANOVA)  design  in  Statistix  (Statistix  8 ©, Analytical  Software,  Tallahassee,  FL, U.S.A.).   All possible  interactions  were  included  and  seasonal  samples were  included  as  replicates.    All  three  factors  were treated as fixed.   Where required data were  log (x + 1) transformed  in order to  normalise the data.  To visualise differences in community structure between the two  regions and between productivity and disturbance treatments, we carried out a non‐ metric multi  dimensional  scaling  (NMDS)  ordination  on  log  (x  +  1)  data  using  Bray‐ Curtis  similarity  and  tested  for  significance  using  analysis  of  similarities  (ANOSIM)  (Clarke 1993) in Primer v6 (Clarke and Gorley 2006).    Regression  analysis  was  used  to  explore  relationships  between  productivity,  disturbance and diversity also using Statistix.   Akaike’s  Information Criterion  (Akaike  1974) was used  to determine  the best  fitting curve  to  the data when  two models  fit  the data equally.   This method  takes  into  account  goodness of  fit  statistics  and  the  number of parameters  involved  in  the  fitting of  the model  to select  the model most  preferred.   We then explored the fit of our data to three main models (and two sub‐ models) of diversity in relation to productivity and disturbance.  These models are the  Dynamic Equilibrium Model  (Huston 1979, 1994) modified by Kondoh  to account  for  meta‐population  dynamics  (Kondoh  2001),  the  Death  (Death  2002)  productivity‐ disturbance‐diversity model, and a model suggested by Tonkin et al.  (In prep.) which  adds to that of Death (2002) by including a quadratic factor between disturbance and  diversity.  The five models tested using multiple regression are of the following form:  1. DEM  S = b0 + b1D + b2C + b3D 2+ b4C 2+b5DxC  2. Death model without interaction between productivity and disturbance  S = b0 + b1D + b2[ln(C)]  3. Death model with interaction between productivity and disturbance  S = b0 + b1D + b2[ln(C)] + b3DxC  4. Tonkin model without interaction between productivity and disturbance  S = b0 + b1D + b2[ln(C)] + b3D 2  Chapter 3: Productivity, disturbance and diversity between landscapes      61    5. Tonkin model with interaction between productivity and disturbance  S = b0 + b1D + b2[ln(C)] + b3D 2+b4DxC    where D = disturbance (percent tracer particle movement), C = chlorophyll a (µg cm‐2),  S = number of  taxa, and b0 – b5 =  regression coefficients.    In models 1, 3 and 5  the  interaction between productivity and disturbance has been  included.   Significance of  this term  indicates whether the two  interact to affect diversity.   That  is, whether the  multiplicative  effects  of  productivity  and  disturbance  is  stronger  than  the  simple  o regression models.   addition of the tw   Results  Physical characteristics and disturbance  Tongariro streams were on average slightly deeper, faster and wider than Hawke’s Bay  streams  (Table  1).   Hawke’s Bay  streams had higher  conductivity  (average: Hawke’s  Bay = 137.5 µS cm‐1, Tongariro = 98.8 µS cm‐1) and temperature (average: Hawke’s Bay  =  14  ˚C,  Tongariro  =  8.8  ˚C)  than  Tongariro  streams.    Little  differences  in  physicochemical  variables  were  evident  between  both  a  priori  high  and  low  productivity groups and stable and unstable groups (Table 1).    There was on average only 14.6 % movement at stable sites compared to 45.8  %  at  unstable  sites  (F1, 59  =  12.49,  P  =  0.0008).    Hawke’s  Bay  streams  had  greater  overall movement  (41.1 %)  compared  to  Tongariro  streams  (21 %)  (F1, 59  =  4.5,  P  =  0.038).  There was no difference in artificial substrate movement between season (F3,  59 = 1.37, P = 0.26).   Overall, measurements of  substrate movement  correlated well  with  the  bottom  component  of  the  Pfankuch  index  (r  =  0.57).    Bed movement will  hereafter be used as the measure of disturbance for the remainder of the analysis.      Periphyton  Of the 16 sites and 63 sets of tiles, 32 sets of tiles remained for assessment, the others  either being washed away or shaded by macrophytes.   Periphyton biomass on stones  was  highly  correlated  to  periphyton  biomass  on  tiles  (r  =  0.86,  P  <  0.0001).   We  Chapter 3: Productivity, disturbance and diversity between landscapes      62    therefore used chlorophyll a from stones as a surrogate for productivity.   Periphyton  biomass,  assessed  as  chlorophyll a,  ranged  from 0.01 µg  cm‐2  to 12.2 µg  cm‐2  at  all  sites.  Periphyton biomass, not surprisingly, was higher at high (1.46 µg cm‐2) than low  productivity sites (0.62 µg cm‐2) (F1, 61 = 5.28, P = 0.025).   Overall periphyton biomass  did not differ between regions  (F1, 59 = 0.15, P = 0.7).   However, periphyton biomass  was significantly higher at stable streams (1.40 µg cm‐2) than unstable streams (0.70 µg  cm‐2)  (F1, 59 = 5.32, P = 0.02).   There was no  interaction between region and stability  (F1,  59  =  0.28,  P  =  0.6)  (Fig.  2).    Periphyton  biomass  declined  with  increasing  bed  disturbance (Fig. 2; Table 2).  This trend was stronger at Hawke’s Bay sites compared to  Tongariro sites but the trend was the same (Fig. 2).    Table 1:  Mean physicochemical characteristics and a priori selected stability and productivity groups for the 16 study streams collected on four occasions  between February 2008 and July 2009 in the North Island of New Zealand.  HB = Hawke’s Bay (non‐pristine), NP = Tongariro (pristine), alt. = altitude, cond. =  conductivity, temp. = temperature, prod. = productivity.    Site      Location      Alt.  (m)    Mean  width  (m)  Mean  depth  (cm)  Mean  velocity  (m s‐1)  Mean  cond.  (µS cm‐1)  Mean  temp.  (˚C)  Mean  % moved    Stability  Prod.  HB1  Makaroro River @ Makaroro Rd  312  5.7  17  0.3  93.3  12.7  95  Unstable  Low  HB2  Waipawa River @ Wakarara Rd  315  6.7  19.4  0.48  96.7  12.8  76.7  Unstable  Low  HB3  Spring creek @ Swamp Rd  146  4.7  43.6  1.52  170  13.7  3.3  Stable  High  HB4  Spring creek @ Lyndsay Rd  134  3.8  22  1.17  230  14.6  40  Stable  High  HB5  Makaretu River @ Burnside Rd  216  5.3  15.8  0.87  96.7  12.9  60  Unstable  High  HB6  Tukipo River @ Burnside Rd  191  7  15  0.2  150  15  30  Unstable  High  HB7  Spring 1 @ Ongaonga Waipuk. Rd  143  1.5  15.2  0.42  153.3  14.8  13.3  Stable  Low  HB8  Spring 2 @ Ongaonga Waipuk. Rd  139  3.2  75.6  0.07  110  15.5  0  Stable  Low  NP1  Wahianoa stream u/s intake  934  6.1  25.6  0.97  70  9.4  50  Unstable  Low  NP2  Unnamed Karioi forest stream  935  3.5  22.6  0.96  139.7  8.6  30  Stable  Low  NP3  Orautoha stream @ middle rd  712  2.5  24.4  0.55  92.7  8.2  8.3  Stable  Low  NP4  Whakapapaiti @ SH4  859  15.8  28.2  0.98  106  8.1  28.3  Unstable  High  NP5  Te Piripi stream @SH1  993  2.2  16.6  0.67  67.7  9.1  10  Stable  High  NP6  Mangatoetoenui @ SH1  971  9.4  44.2  0.96  133  9.5  25  Unstable  Low  NP7  Oturere Stream SH1  809  9.4  42.4  0.86  110.7  8.6  8.3  Unstable  High  NP8  Poutu Stream  518  7.7  43.6  1.05  70.3  8.9  8.3  Stable  High      Chapter 3: Productivity, disturbance and diversity between landscapes      64      Figure 2: Mean  chlorophyll a  concentrations  (± 1  S.E.)  (a)  for  sites  in each a priori  selected  stability group in both regions and (b) as a function of percent stone movement collected in 16  streams, North  Island, New Zealand, February 2008‐July 2009. Black bars/circles are pristine  (Tongariro) and open bars/circles are non‐pristine (Hawke’s Bay).  Solid line = all sites, dashed  line = non‐pristine sites and dotted line = pristine sites.  See text for ANOVA results and Table 2  for regression results.      Community composition  One hundred and eleven taxa were collected from all sites in this study.  Although the  Hawke’s Bay sites had more  than  twice  the number of collected animals  (93,609)  to  that of the Tongariro sites (40,570), 23 more taxa were collected from Tongariro (97)  than  in Hawke’s Bay  (74).    Insect  taxa dominated  invertebrate  communities  in both  regions.   Ephemeroptera  (16.5 %), Trichoptera  (28.8 %) and Diptera  (28 %) made up  73.4 % of all animals collected.   This structure was slightly different between the two  regions with 67.8 %  insects  in Hawke’s Bay  and 86.2 %  in  Tongariro. Unstable  sites  were  dominated  by  the mayfly  Deleatidium  spp.  averaging  47.3 %  at  unstable‐low  productivity  sites  and  21.1  %  at  unstable‐high  productivity  sites.    Stable‐low  Chapter 3: Productivity, disturbance and diversity between landscapes      65    productivity sites were dominated by molluscan taxa  (25.8 %) and oligochaetes  (18.5  %).      Table  2:  Results  of  regression  analysis  of  chlorophyll  a  and  three  invertebrate  community  metrics against (a) % stone movement and (b) chlorophyll a collected in 16 North Island, New  Zealand streams on  four occasions between February 2008 and  July 2009.   D.f. = degrees of  freedom.  AIC = Akaike’s information criterion where lower numbers represent a better model.         D.f.  F (AIC)  P  r2 Equation  a) % moved  Chlorophyll a  1, 61  20.1 <0.0001 0.25 y = 0.32 ‐ 0.0028x         Pristine  1, 30  6.81 0.014 0.19 y = 0.29 ‐ 0.003x         Non‐pristine 1, 29  19.28 0.0001 0.4 y = 0.37 ‐ 0.003x  No. of individuals 1, 61  19.49 (‐31.01) <0.0001 0.24 y = 2.53 ‐ 0.007x                quadratic  2, 60  10.55 (‐29.67)  0.0001  0.26  y = 2.49 + 0.002x ‐ 0.00009x2        Pristine  1, 30  23.04 <0.0001 0.43 y = 2.33 ‐ 0.012x         Non‐pristine 1, 29  19.65 0.0001 0.4 y = 2.76 ‐ 0.007x  No. of taxa  1, 61  33.01 (‐93.91) <0.0001 0.35 y = 1.26 ‐ 0.003x                quadratic  2, 60  19.71 (‐93.90)  <0.0001  0.4  y = 1.23 + 0.002x ‐ 0.00005x2        Pristine  1, 30  18.60 0.0002 0.38 y = 1.25 ‐ 0.0045x         Non‐pristine 1, 29  37.86 <0.0001 0.57 y = 1.27 ‐ 0.0024x  Rarefied richness 1, 56  5.92 0.018 0.1 y = 19.6 ‐ 0.04x         Pristine  1, 26  0.04 0.84 0.002 Non‐significant         Non‐pristine 1, 28  5.15 0.03 0.16 y = 17.55 ‐ 0.03x  b) Periphyton biomass  No. of individuals 1, 61  67.98 <0.0001 0.53 y = 2.98 + 0.32∙ln(x)        Pristine  1, 30  34.66 <0.0001 0.53 y = 2.83 + 0.35∙ln(x)        Non‐pristine 1, 29  84.44 <0.0001 0.74 y = 3.14 + 0.29∙ln(x) No. of taxa  1, 61  48.64 <0.0001 0.44 y = 1.38 + 0.1∙ln(x)         Pristine  1, 30  26.67 <0.0001 0.47 y = 1.44 + 0.13∙ln(x)        Non‐pristine 1, 29  32.96 <0.0001 0.53 y = 1.32 + 0.07∙ln(x) Rarefied richness 1, 56  0.09 0.76 0.002 Non‐significant         Pristine  1, 26  0.02 0.9 0.0006 Non‐significant         Non‐pristine 1, 28  0.98 0.33 0.03 Non‐significant      NMDS  (Fig. 3) and ANOSIM confirmed community structure differed between  the  two  regions  (R  =  0.42,  P  =  0.001,  Fig.  3),  and  using  presence/absence  data  suggested  it was  not  only  abundances  that were  different  but  also  composition  of  species (R = 0.47, P = 0.001).  The greatest variation in densities between communities  occurred within the pristine unstable sites (avg. similarity: 35.9) whereas non‐pristine  unstable sites were more similar  (avg. similarity: 48.7).   Non‐pristine stable sites had  the most similar communities (avg. similarity: 52.6; pristine stable avg. similarity: 43.2).   No single taxon contributed more than 5.6 % to the difference between communities  between the regions.   Of the most  important taxa, Pycnocentrodes aeris, Deleatidium  Chapter 3: Productivity, disturbance and diversity between landscapes      66    spp., Oligochaeta, Potamopyrgus antipodarum and Paracalliope  fluviatilis were more  prevalent  in  Hawke’s  Bay  streams  whereas  Maoridiamesa  spp.,  two  orthoclad  chironomids,  Zelandoperla  spp.  and Hydora  spp. were more  abundant  in  Tongariro.   Although the difference in community structure was greatest between regions, it also  differed  between  stable  and  unstable  sites  (R  =  0.16,  P  =  0.001,  Fig.  3)  and  less  so  between  low  and  high  productivity  sites  (R  =  0.06,  P  =  0.017,  Fig.  3).    However,  community dissimilarities with both  stability  and productivity were due primarily  to  differences  in  relative  abundances  of  dominant  taxa  rather  than  changes  in  composition.   Deleatidium spp., a species of Orthocladiinae, Maoridiamesa spp., and  the  cased  caddisfly  Pycnocentrodes  aeris  contributed  to  16.4  %  of  the  difference  between stable and unstable sites and 17.7 % between low and high productivity sites.   All of these taxa were higher at high productivity sites and, barring Deleatidium spp.,  higher at stable sites.        Figure 3: Non‐metric multidimensional scaling  (NMDS) ordination on  log  (x + 1)  transformed  data for 16 streams collected on four occasions between February 2008 and July 2009  in the  North Island of New Zealand.  Closed circles are pristine sites (Tongariro Region), open circles  are  non‐pristine  sites  (Hawke’s  Bay  region).    Unstable/low  productivity:  upright  triangles;  Unstable/high  productivity:  inverted  triangles;  Stable/low  productivity:  squares;  Stable/high  productivity: circles.      Chapter 3: Productivity, disturbance and diversity between landscapes      67    Density and diversity  The density of invertebrates per 0.1 m2 was significantly higher at Hawke’s Bay streams  with  603.93  individuals  0.1  m‐2  opposed  to  Tongariro’s  253.58  individuals  0.1  m‐2  (Table 3).  Density was significantly higher at stable than unstable streams, with 578.63  individuals 0.1 m‐2 at stable streams and 278.1 individuals 0.1 m‐2 at unstable streams  (Table 3).   There was no  interaction between stability and  region  (Table 3).   Density  was  significantly  higher  at  high  productivity  sites with  515.8  individuals  and  333.2  individuals  at  low  productivity  sites  (Table  3).    There  was  no  interaction  between  productivity and stability nor was there a three way interaction including region (Table  3).  Density increased logarithmically with increasing periphyton biomass and declined  with increasing bed movement (Fig. 4; Table 2).  These relationships were the same for  both regions.  However when fitted individually, slight differences in the curves meant  each model explained a greater proportion of the variation in the data (Fig. 4; Table 2).      Table 3: Three‐way ANOVA results  testing  for differences  in mean number of animals, mean  number  of  taxa,  and  rarefied  number  of  taxa  between  region  and  a  priori  stability  and  productivity  groups  collected  in  16  North  Island,  New  Zealand  streams  on  four  occasions  between February 2008 and July 2009.       No. of animals  No. of taxa  Rarefied  richness  (ES224)     F P F P F  P Stability  18.19 0.0001 28.81 <0.0001  5.89  0.018 Productivity  13.69 0.0005 15.32 0.0003  0.28  0.6 Region  18.93 0.0001 0.17 0.68 11.58  0.001 Stability x productivity  7.79 0.007 11.31 0.001  3.91  0.053 Stability x region 0.38 0.54 1.95 0.17 2.37  0.13 Productivity x region  3.53 0.065 0.96 0.33 1.59  0.21 Stability x productivity x region  0.81 0.37 7.05 0.01 10.51  0.002     The mean number of taxa per sample at each site (richness) was slightly higher  on  average  in  Tongariro  streams with  15.49  taxa  0.1 m‐2  and  14.57  taxa  0.1 m‐2  at  Hawke’s  Bay  streams;  however,  this  was  not  a  significant  difference  (Table  3).   Richness was higher at stable sites with 18.11 taxa 0.1 m‐2 and 12.07 taxa 0.1 m‐2 at  unstable sites (Table 3).  Richness was higher at high productivity sites with 17 taxa and  13 at  low productivity  sites  (Table 3).   Productivity and  stability  interacted  to effect  richness and there was also a three‐way interaction between productivity, stability and  Chapter 3: Productivity, disturbance and diversity between landscapes      68    region  (Table  3).    Taxonomic  richness  increased  logarithmically  with  increasing  periphyton biomass and declined with decreasing bed stability (Fig. 5; Table 2).  These  trends were  consistent  for  both  regions  but  the  fit  of  both models was  greater  at  Hawke’s Bay sites (Fig. 5; Table 2).    Rarefied  taxa  richness,  pooled  for  224  individuals  was  higher  at  Tongariro  streams with  20.74  taxa  at  these  streams  and  16.41  taxa  at  Hawke’s  Bay  streams  (Table 3).   Rarefied  richness was higher at  stable  sites  (19.71)  than at unstable  sites  (16.89), but there was no difference between high and low productivity groups (Table  3).    There  was  no  interaction  between  region  and  stability,  but  productivity  and  stability did  interact with  region  (Table 3).   Rarefied  richness declined gradually with  increasing bed movement (Fig. 6; Table 2).  When fitting to each individual region, no  relationship was  found  between  bed movement  and  rarefied  richness  at  Tongariro  sites  (Fig.  6;  Table  2).    There was  no  relationship  between  periphyton  biomass  and  rarefied richness at all sites (Fig. 6; Table 2).        Chapter 3: Productivity, disturbance and diversity between landscapes      69      Figure 4: Mean number of animals (± 1 S.E.) collected in 16 North Island, New Zealand streams  on four occasions between February 2008 and July 2009.  (a) For each a priori selected stability  and productivity group, (b) as a function of mean chlorophyll a and (c) as a function of percent  stone movement.   Black bars/circles  are pristine  (Tongariro) and open bars/circles are non‐ pristine (Hawke’s Bay).  Solid line = all sites, dashed line = non‐pristine sites and dotted line =  pristine sites.  See Table 3 for ANOVA results and Table 2 for regression results.    Chapter 3: Productivity, disturbance and diversity between landscapes      70      Figure 5: Mean number of taxa (± 1 S.E.) collected in 16 North Island, New Zealand streams on  four occasions between February 2008 and  July 2009.   (a) For each a priori selected stability  and productivity group, (b) as a function of mean chlorophyll a and (c) as a function of percent  stone movement.       Black bars/circles are pristine  (Tongariro) and open bars/circles are non‐ pristine (Hawke’s Bay).  Solid line = all sites, dashed line = non‐pristine sites and dotted line =  pristine sites.  See Table 3 for ANOVA results and Table 2 for regression results.        Chapter 3: Productivity, disturbance and diversity between landscapes      71      Figure  6:  (a) Mean  rarefied  richness  (± 1  S.E.)  calculated  for 224  individuals  collected  in 16  North  Island, New Zealand streams on  four occasions between February 2008 and  July 2009  for  each  a  priori  selected  stability  and  productivity  group.    (b)  Pooled  rarefied  richness  calculated  for  224  individuals  as  a  function  of mean  chlorophyll  a  and  (c)  as  a  function  of  percent stone movement.  Black bars/circles are pristine (Tongariro) and open bars/circles are  non‐pristine  (Hawke’s Bay).   Solid  line = all sites, dashed  line = non‐pristine sites and dotted  line = pristine sites.  See Table 3 for ANOVA results and Table 2 for regression results.    Chapter 3: Productivity, disturbance and diversity between landscapes      72    Model fitting  All five models explained a  large proportion of the variation  in taxonomic richness at  all sites  (Fig. 7; Table 4; Table 5).   The DEM explained 50.9 % of  the variation  in  the  data  for  the number of  taxa  (Fig. 7; Table 4).   The Tonkin model explained  the most  variation  in  the  data  with  52.4  %  when  the  interaction  between  productivity  and  disturbance was  included  and 51.9 % without  the  interaction  (Fig. 7;  Table 4).    The  Death  model  explained  47.9  %  of  the  variation  with  and  without  the  interaction  between productivity and disturbance included (Fig. 7; Table 4).  Akaike’s information  criterion  shows  that  the  Tonkin  and Death models without  interaction  are  the  best  models based on a trade‐off between simplicity and fit statistics (Fig. 7; Table 4).  The  difference between  the  two AIC values  is negligible and  thus we  can  consider  these  two equally well fitting.  Our model fitting shows that there is no interaction between  the way productivity and disturbance effect  richness,  thus  the effect of productivity  and  disturbance  on  diversity  in  these  streams  is  additive  rather  than multiplicative.   The only model of the  five to have every coefficient significantly affecting richness  is  the Death model without  interaction  (Fig. 7;  Table 5).    The productivity  coefficients  were  significant  throughout  all  five models whereas  the only  significant disturbance  coefficients were the disturbance2 coefficients.  Although the DEM explains more than  the Death model, due to the complexity of the model it results as the worst model for  the  data  (Fig.  7;  Table  4).   Although  the Death model  explains  the  least  amount  of  variation in the data, the difference between the amount of variation explained by all  five models is small (0.48 < r2 < 0.52) (Fig. 7; Table 4).    When  fitting  the  models  to  each  individual  region,  the  Death  and  Tonkin  models without the interaction term were the most applicable in both regions (Fig. 7;  Table 4).  However, model fit was consistently better in the non‐pristine sites than the  pristine sites (Fig. 7; Table 4).  Although the DEM did explain a large proportion of the  variation, AIC  indicated  it was  the  least  likely model  for  these data  (Fig. 7; Table 4).   The pristine  region  responded differently  to  the  three models with  the productivity‐ disturbance interaction (Fig. 7).   The two models with the  interaction and a quadratic  relationship between disturbance and diversity suggested diversity at the pristine sites  would decline significantly at high levels of disturbance and productivity (Fig. 7).    Chapter 3: Productivity, disturbance and diversity between landscapes      73      Figure 7: Model fit for mean number of taxa collected from 16 North Island, New Zealand sites  on  four occasions between February 2008 and  July 2009 with all sites  included, pristine only  (Tongariro) and non‐pristine only  (Hawke’s Bay)  sites.  (a) DEM,  (b) Death model – no P  x D  (productivity x disturbance)  interaction,  (c) Death model – with P x D  interaction,  (d) Tonkin  model – no P x D  interaction, and (e) Tonkin model – with P x D  interaction.   See Table 4 for  model fit results and Table 5 for model coefficients.         Chapter 3: Productivity, disturbance and diversity between landscapes      74    Table 4: Model  fit results  for  five models predicting mean number of taxa collected  from 16  North Island, New Zealand sites on four occasions between February 2008 and July 2009 with  (a) all sites included, (b) pristine only (Tongariro) and (c) non‐pristine only (Hawke’s Bay) sites.   With  interaction  indicates  the productivity  x disturbance  interaction  term  is  included  in  the  model.   RMSE =  root mean  square error.   AIC = Akaike’s  information  criterion where  lower  numbers represent a better model.       D.f.  F  P  RMSE  r2 AIC  a) All sites  DEM  5, 57  11.82  <0.0001  0.15  0.51  ‐93.55  Death  2, 60  27.55  <0.0001  0.15  0.48  ‐97.9  Death (with interaction)  3, 59  18.06  <0.0001  0.15  0.48  ‐95.91  Tonkin  3, 59  21.17  <0.0001  0.14  0.52  ‐98.07  Tonkin (with interaction)  4, 58  15.98  <0.0001  0.14  0.52  ‐96.41  b) Pristine  DEM  5, 26  7.71  0.0001  0.17  0.6  ‐37.68  Death  2, 29  14.89  <0.0001  0.18  0.51  ‐40.86  Death (with interaction)  3, 28  9.85  0.0001  0.18  0.51  ‐39.05  Tonkin  3, 28  12.08  <0.0001  0.17  0.56  ‐40.58  Tonkin (with interaction)  4, 27  9.4  <0.0001  0.17  0.58  ‐39.17  c) Non‐pristine  DEM  5, 25  11.58  <0.0001  0.08  0.7  ‐55.53  Death  2, 28  24.71  <0.0001  0.09  0.64  ‐59.08  Death (with interaction)  3, 27  15.97  <0.0001  0.09  0.64  ‐57.13  Tonkin  3, 27  21.49  <0.0001  0.08  0.7  ‐59.82  Tonkin (with interaction)  4, 26  16.51  <0.0001  0.08  0.72  ‐58.41          Table 5: Model coefficients results  for  five models predicting mean number of taxa collected  from 16 North  Island, New Zealand sites on  four occasions  between February 2008 and July 2009 with  all sites included, pristine only (Tongariro) and non‐pristine only (Hawke’s Bay) sites.  With interaction indicates  the productivity x disturbance interaction term is included in the model.  D = disturbance and P = productivity.       All sites                  Pristine    Non pristine           Coefficient (S.E.)  t  P     Coefficient (S.E.)  t  P     Coefficient (S.E.)  t  P  DEM  y0  1.07 (0.06)  18.14  <0.0001  1.03 (0.08)  13.13  <0.0001  1.11 (0.08)  13.28  <0.0001  D  0.004 (0.003)  1.55  0.13  0.01 (0.006)  1.75  0.092  0.004 (0.003)  1.49  0.148  P  0.79 (0.28)  2.82  0.007  1.06 (0.42)  2.54  0.017  0.65 (0.36)  1.79  0.086  D2           ‐   ‐     ‐     ‐     ‐     ‐     ‐     ‐     ‐     ‐     ‐     ‐     ‐     ‐           ‐0.00006 (0.00002) ‐2.44  0.017 ‐0.0001 (0.00001) ‐2.38  0.025 ‐0.0001 (0.00002) 2.4  0.024  P2 0.6 (0.28) 2.17  0.034 ‐0.76 (0.4) 1.91  0.067 ‐0.65 (0.38) 1.73  0.095  DxP  0.0001 (0.005)  0.02  0.98 ‐0.03 (0.02) 1.02  0.32 ‐0.0004 (0.004) 0.12  0.91  Death  y0  1.36 (0.04)  37.99  <0.0001  1.41 (0.07)  20.63  <0.0001  1.31 (0.03)  46.71  <0.0001  D ‐0.001 (0.0006) 2.01  0.049 ‐0.002 (0.001) 1.46  0.16 ‐0.002 (0.001) 2.87  0.008  P  0.07 (0.02)  3.83  0.0003  0.1 (0.04)  2.7  0.011  0.04 (0.02)  2.36  0.025  Death (with interaction)  y0  1.36 (0.05)  29.78  <0.0001  1.38 (0.08)  17.91  <0.0001  1.3 (0.04)  30.93  <0.0001  D ‐0.001 (0.0009) 1.31  0.19 ‐0.003 (0.002) 1.55  0.13 ‐0.002 (0.001) 1.78  0.087  P  0.08 (0.03)  2.92  0.005  0.09 (0.04)  2.25  0.033  0.03 (0.03)  1.13  0.27  DxP ‐0.0003 (0.005) 0.06  0.95  0.01 (0.02)  0.62  0.54  0.001 (0.004)  0.3  0.77  Tonkin  y0  1.34 (0.04)  36.88  <0.0001  1.37 (0.07)  19.9  <0.0001  1.25 (0.03)  40.64  <0.0001  D  0.004 (0.002)  1.56  0.12  0.005 (0.004)  1.27  0.22  0.002 (0.03)  0.98  0.33  P  0.07 (0.02)  3.86  0.0003  0.09 (0.04)  2.49  0.019 ‐0.002 (0.07) 0.02  0.98  D2 ‐0.00005 (0.00002) ‐2.21  0.031 ‐0.0001 (0.00004) ‐1.92  0.065 ‐0.00004 (0.00002) ‐2.13  0.042    Chapter 3: Productivity, disturbance and diversity between landscapes      76                 Table 5 continued....     All sites       Pristine    Non pristine           Coefficient (S.E.)  t  P     Coefficient (S.E.)  t  P     Coefficient (S.E.)  t  P  Tonkin (with interaction)  y0  1.36 (0.04)  30.8  <0.0001  1.39 (0.07)  19.07  <0.0001  1.32 (0.04)  34.31  <0.0001  D  0.005 (0.003)  1.76  0.083  0.01 (0.01)  1.63  0.115  0.01 (0.003)  1.92  0.067  P  0.09 (0.03)  3.43  0.001  0.1 (0.04)  2.72  0.011  0.07 (0.03)  2.31  0.029  D2               ‐     ‐       ‐ ‐0.00006 (0.00002) ‐2.36  0.022 ‐0.0001 (0.0001) ‐2.11  0.045 ‐0.0001 (0.00002) ‐2.68  0.013  DxP ‐0.005 (0.005) 0.84  0.4    ‐0.03 (0.03) 1.08  0.29 0.29 1.08   ‐0.005 (0.005)               Chapter 3: Productivity, disturbance and diversity between landscapes      Discussion  77    Diversity  of  invertebrate  communities  responded  strongly  to  both  productivity  and  disturbance in this study.  We found evidence to support Death’s (2002) productivity‐ disturbance‐diversity model  in  these  streams.    Our  results  suggest  Huston’s  (1979,  1994) DEM  is not  applicable  in  these  streams.   However we  also  found  support  for  Tonkin  et  al.’s  (In  prep.) modification  of  the Death model  of  diversity  to  include  a  quadratic disturbance‐diversity relationship.   There was  little difference  in the way  in  which productivity and disturbance affected diversity between  the pristine and non‐ pristine  regions.   Curiously, not one of  the models exhibited a  significant  interactive  effect  of  productivity  and  disturbance.      There  was  no  evidence  of  competitive  exclusion  in  these  streams  with  diversity  increasing  and  tapering  with  increasing  productivity.    This  is  a  phenomenon  widely  found  in  lotic  environments  with  productivity setting the upper limit to carrying capacity in streams (Death 2010).      Regional differences  Although  the  composition  and density of  the benthic  communities differed  strongly  between the pristine (Tongariro) and non‐pristine (Hawke’s Bay) regions, the effects of  productivity and disturbance on diversity remained similar between  the  two regions.   However,  the  productivity‐disturbance‐diversity  relationship  was  stronger  at  non‐ pristine than pristine sites.   Taxonomic richness was not affected by region but there  was a significant three‐way effect of productivity, disturbance and region.  Changes in  land use from forested to pastoral grazing can have several effects such as changes to  hydrology,  sediment,  channel morphology,  light  availability  (and  the  resulting  food  base shift) and temperature (Quinn 2000, Allan 2004).  Changes in catchment land use  can  lead  to  flow  on  effects  such  as  severe  changes  to  flood  regimes  resulting  from  urbanisation (Walsh et al. 2005) and increased peak flows during flood events resulting  from  deforestation  (Rowe  et  al.  1997).   When  investigating  a  one‐in‐28‐year  flood,  Collier  and  Quinn  (2003)  found  flood  disturbance  can  lead  to  differential  effects  between  pastoral  and  forested  streams.    They  found  that  although  richness  and  density at the pasture site exhibited a delayed response compared to the forested site,  Chapter 3: Productivity, disturbance and diversity between landscapes      78    community  structure  at  the  pasture  site  alone was  destabilised  by  the  flood.    They  suggest that the differential response of the two sites was due to the presence of an  underlying  press  disturbance  (land  use)  at  the  pasture  site,  but without  replication  they could not confirm whether  it was simply  individual site differences.    In contrast,  communities  in  the unstable‐pristine  sites  in our  study were  the  least  similar  rather  than pastoral  sites.   The  large  scale of disturbance  in  the study of Collier and Quinn  (2003)  which  left  few  colonists  available  post‐disturbance,  makes  comparing  our  results with those found in their study impractical.    Not only can changes  in  land use alter the composition of communities within  these  landscapes  (Harding et al. 1998, Allan 2004, Tscharntke et al. 2005), but  local  inter‐species  interactions are  limited by  larger scale processes such as the availability  of  colonists  and  dispersal  abilities.    Thus,  although  the  colonist  pool  is  likely  to  be  different between the two regions, the response to productivity and disturbance was  similar between the pristine and non‐pristine regions.  Several studies have suggested  that  local diversity‐productivity relationships could be heavily  influenced by dispersal  and  colonisation  processes  at  a  regional  scale  which  may  in  turn  generate  local  processes  such  as  competition  (Tilman  1999a,  Loreau  et  al.  2001,  Cardinale  et  al.  2005).    The only exception  to  this was  the differential effects of bed movement on  rarefied richness between the two regions with no link evident at pristine sites.  Being  part  of  a  braided  river  deposit  system,  substrate  in  the  Hawke’s  Bay  was  less  embedded  and  loose  and  thus  diversity  may  have  been  affected  more  than  the  measured movement indicates.      Model fitting  Although  the  Death  (2002)  productivity‐disturbance‐diversity  model  explained  the  least  amount  of  variation  in  the  data,  it  was  also  the most  simplistic  and  thus  is  favourable to the more complex DEM.  There was only a small difference in predictive  ability between all three models.   Moreover, the model without the  interaction term  was the only model with all coefficients significant.  While our data fit slightly better to  the  model  of  Tonkin  et  al.  (In  prep.),  the  difference  in  strength  of  fit  was  negligible.  The only difference between the models put forward by Death (2002) and  Chapter 3: Productivity, disturbance and diversity between landscapes      79    Tonkin et al. (In prep.)  is the relationship between disturbance and diversity: Death’s  model advocates for a linear decline in diversity rather than the quadratic of Tonkin et  al..  If the quadratic relationship between disturbance and diversity was of a unimodal  form in the model of Tonkin et al. (In prep.) we could postulate competitive exclusion  was occurring but this was not the case.  Similarly, there was no decline in diversity at  extremely stable sites as would be the case with the DEM.  Although the IDH is one of  the  most  widely  researched  patterns  in  ecology  (Connell  1978),  the  unimodal  relationship between disturbance and diversity  is  rare  in nature  (Mackey and Currie  2001).  Likewise, there has been limited support of this unimodal trend in lotic systems  but  Townsend  et  al.  (1997)  found  diversity  peaked  at  intermediate  levels  of  disturbance  in  South  Island  New  Zealand  streams.    Svensson  et  al.  (2007)  found  support for the IDH when testing the DEM but found no change in the effect of the IDH  through multiple  levels  of  productivity.   Death  and Winterbourn  (1995)  found  that  stream  invertebrate  diversity  declined with  disturbance  irrespective  of  productivity  levels.    Like  Tonkin  et  al.’s  (In  prep.)  study  on  the  disturbance‐productivity‐diversity  relationship, this study is also limited by a low range of sampled productivity, however  there  is  still  a  clear  linear  decline  in  periphyton  biomass with  disturbance.  Scholes  (2005) mentioned  the  importance  of  sampling  adequate  ranges  of  productivity  and  disturbance  in  tests  of  this  nature  in  order  to  fully  evaluate  the  effects  operating.   Death  (2002)  also  found  a  clear  effect  of  disturbance  within  a  small  range  of  periphyton biomass.  If, however, higher levels of periphyton biomass were present we  might speculate that diversity may have declined at greater levels as has been found in  numerous other cases  (Rosenzweig 1995, Waide et al. 1999, Mittelbach et al. 2001).   However,  this  is  dependent  on  competitive  exclusion  operating which  as  previously  discussed  is not  likely  in  these  streams.   Tonkin et al.  (In prep.) also  found  this  log‐ linear increase but this was only evident at open canopy streams within the Tongariro  National  Park  suggesting  productivity  is  setting  the  upper  limit  to  diversity  in  these  streams.    As hypothesised, we found  little evidence to suggest the DEM applies  in these  streams.  The DEM has received limited support in lotic systems (but see Cardinale et  Chapter 3: Productivity, disturbance and diversity between landscapes      80    al.  2006).  The  lotic  environment  presents  a  highly  open  system  which  leads  to  community dynamics controlled by  immigration and emigration through  invertebrate  drift and faunas dominated by highly mobile taxa as opposed to many other habitats  being  controlled  by  population  growth  through  reproduction  (Williams  and  Hynes  1976, Brittain and Eikeland 1988, Mackay 1992).  This  is where  the DEM has  limited  applicability  in  stream  communities.  The  DEM  predicts  that  at  high  productivity/resource  levels  the  competitive  dominant  will  monopolise  resources  through increased reproduction.  However in streams, immigration will be greater than  emigration and  thus diversity will simply  increase  (Death 2002).   We  found  this here  with diversity  increasing to a point and then  levelling off with  increasing productivity.   Essentially high resource levels are lowering the immigration to emigration rate rather  than causing competitive exclusion as would be the case  if the DEM were operating.   As discussed by Death  (2002), productivity  is  setting  the upper  limit  for  the carrying  capacity of the stream, rather than  leading to competitively dominant taxa excluding  other less competitive taxa.  For the DEM to be applicable to these streams, extinction  rates would have to outweigh  immigration rates during recolonisation (Petraitis et al.  1989).    Disturbance not only acts by direct removal of invertebrates but by the removal  of periphyton, a major component of the food web and potentially the main food base  of  stream  food  webs  (Robinson  and  Minshall  1986,  Death  2002,  Death  and  Zimmermann 2005).  In fact, this removal of periphyton may be more  important than  the  removal  of  invertebrates  themselves  as  the  recolonisation  of  invertebrates  is  dependent on  the resources present at a site.   We have  found  that productivity and  disturbance do  indeed explain a  large portion of  the variation  in diversity, however,  like Tonkin et al.  (In prep.) we did not  find any evidence  that  these  two  factors are  interacting.  These interactive effects of productivity and disturbance have been found  in streams (e.g. Cardinale et al. 2006), however, what we found  is that the effects of  productivity  and  disturbance were  additive  rather  than multiplicative.    Perhaps  the  lack of interaction is indicative of the fact that models predicting interactive effects of  productivity  and  disturbance  are  reliant  on  disturbance  disrupting  biotic  processes  such as competitive exclusion which we don’t believe are occurring here.  Although the  Chapter 3: Productivity, disturbance and diversity between landscapes      81    inclusion of  the  interaction  term between productivity and disturbance  in  the Death  and Tonkin models  revealed  interesting  changes,  they were not  significant and only  improved model fit moderately, if at all.    Both  the  number  of  taxa  and  the  total  number  of  animals  were  highly  correlated  and  followed  the  same  trends  of  log‐linear  increase  with  increasing  periphyton  and  linear  decline  with  increasing  disturbance.    This  suggests  that  the  number of taxa collected may simply be a function of the number of animals collected.   We  found  that when  density was  accounted  for  in  the  assessment  of  richness,  the  relationship between richness and productivity disappears.  However, rarefied richness  still  declined  with  increasing  disturbance.    Previous  studies  have  found  rarefied  richness  patterns  to  operate  differently  to  simple  richness  measures.    Death  and  Zimmerman  (2005)  found  that  rarefied  richness declined with  increasing disturbance  and  Death  (2002)  found  lower  rarefied  richness  at  more  disturbed  streams.   Conversely, McCabe and Gotelli (2000) found higher rarefied species richness at more  disturbed experimental treatments as opposed to the number of species being  lower  at these treatments.  The lack of response of rarefied richness to productivity suggests  that although resources may be greater at higher productivity sites leading to a higher  density of animals, this does not necessarily mean that the diversity of resources has  increased and opened up new habitat  for new taxa.   Essentially, more resources will  likely lead to more individuals per species while more taxa are likely to appear when a  greater diversity of resources is present rather than a greater volume of resources.      Conclusions  We  have  shown  that  a  model  predicting  a  log‐linear  increase  in  diversity  with  productivity  and  either  a  linear  or  quadratic  decline with  disturbance  to  be  highly  applicable to predicting diversity (Death 2002, Tonkin et al. In prep.).   We found  little  evidence to support the DEM (Huston 1979, 1994), and found no evidence in this study  of competitive exclusion as required by the DEM. The number of taxa did not decline  at higher  levels of periphyton nor did  it at higher  stability  sites.   We  found  that  the  response of diversity to productivity and disturbance was similar between pristine and  non‐pristine  regions  even  though  community  composition  differed  between  the  Chapter 3: Productivity, disturbance and diversity between landscapes      82    regions.    The  main  differences  found  were  the  productivity‐disturbance‐diversity  relationship was stronger at non‐pristine sites and communities at unstable sites were  more variable within a pristine as opposed to anthropogenically  impacted  landscape.   Despite  research  suggesting  disturbance  operates  on  stream  communities  both  through direct impact on animals and their resource supply, we found their effects on  diversity, although strong, to be additive rather than multiplicative.   We suggest both  the Death (2002) model and Tonkin et al. (In prep.) model are applicable to ecological  communities with highly mobile animals irrespective of the landscape they are applied  within.      References    Akaike,  H.  1974.  New  look  at  statistical‐model  identification.  Ieee  Transactions  on  Automatic Control AC19:716‐723.  Allan, J. D. 1995. Stream Ecology: Structure and Function of Running Waters. Chapman  and Hall, London.  Allan,  J. D.  2004.  Landscapes  and  riverscapes:  the  influence  of  land  use  on  stream  ecosystems. Annual Review of Ecological and Evolutionary Systematics 35:257‐ 284.  Brittain,  J.  E.,  and  T.  J.  Eikeland.  1988.  Invertebrate  drift  ‐  a  review.  Hydrobiologia  166:77‐93.  Cardinale, B. J., H. Hillebrand, and D. F. Charles. 2006. Geographic patterns of diversity  in  streams  are  predicted  by  a  multivariate  model  of  disturbance  and  productivity. Journal of Ecology 94:609‐618.  Cardinale, B. J., M. A. Palmer, A. R. Ives, and S. S. Brooks. 2005. Diversity‐productivity  relationships  in streams vary as a  function of  the natural disturbance  regime.  Ecology 86:716‐726.  Clarke,  K.  R.  1993.  Nonparametric  multivariate  analyses  of  changes  in  community  structure. Australian Journal of Ecology 18:117‐143.  Chapter 3: Productivity, disturbance and diversity between landscapes      83    Clarke,  K.  R.,  and  R. N. Gorley.  2006.  PRIMER  v6: User Manual/Tutorial.  PRIMER‐E,  Plymouth.  Collier, K. J., and J. M. Quinn. 2003. Land‐use influences macroinvertebrate community  response following a pulse disturbance. Freshwater Biology 48:1462‐1481.  Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs. Science 199:1302‐ 1310.  Currie, D. J. 1991. Energy and large‐scale patterns of animal‐species and plant‐species  richness. American Naturalist 137:27‐49.  Death, R. G. 2002. Predicting invertebrate diversity from disturbance regimes in forest  streams. Oikos 97:18‐30.  Death, R. G. 2008. Effects of floods on aquatic invertebrate communities.in J. Lancaster  and R. A. Briers, editors. Insects: Challenges to Populations. CAB International,  UK, pp 103‐121.  Death,  R.  G.  2010.  Disturbance  and  riverine  benthic  communities:  What  has  it  contributed  to  general  ecological  theory?  River  Research  and  Applications  26:15‐25.  Death, R. G., and K. J. Collier. 2010. Measuring stream macroinvertebrate responses to  gradients  of  vegetation  cover: when  is  enough  enough?  Freshwater  Biology  55:1447‐1464.  Death,  R. G.,  and M.  J. Winterbourn.  1994.  Environmental  stability  and  community  persistence:  a  multivariate  perspective.  Journal  of  the  North  American  Benthological Society 13:125‐139.  Death,  R.  G.,  and  M.  J.  Winterbourn.  1995.  Diversity  patterns  in  stream  benthic  invertebrate communities:  the  influence of habitat  stability. Ecology 76:1446‐ 1460.  Death,  R.  G.,  and  E. M.  Zimmermann.  2005.  Interaction  between  disturbance  and  primary  productivity  in  determining  stream  invertebrate  diversity.  Oikos  111:392‐402.  Downes, B. J., P. S. Lake, A. Glaister, and J. A. Webb. 1998. Scales and frequencies of  disturbances:  rock  size,  bed  packing  and  variation  among  upland  streams.  Freshwater Biology 40:625‐639.  Chapter 3: Productivity, disturbance and diversity between landscapes      84    Foley, J. A., R. DeFries, G. P. Asner, C. Barford, G. Bonan, S. R. Carpenter, F. S. Chapin,  M. T. Coe, G. C. Daily, H. K. Gibbs, J. H. Helkowski, T. Holloway, E. A. Howard, C.  J.  Kucharik,  C. Monfreda,  J. A.  Patz,  I.  C.  Prentice, N.  Ramankutty,  and  P.  K.  Snyder. 2005. Global consequences of land use. Science 309:570‐574.  Graham, A. A., D. J. McCaughan, and F. S. McKee. 1988. Measurement of surface area  of stones. Hydrobiologia 157:85‐87.  Harding,  J.  S.,  E.  F.  Benfield,  P.  V.  Bolstad,  and  E.  B.  D.  Jones  III.  1998.  Stream  biodiversity: The ghost of  land use past. Proceedings of the National Academy  of Sciences 95:14843‐14847.  Hildrew, A. G., and P. S. Giller. 1994. Patchiness, species  interactions and disturbance  in  the  stream  benthos.  Pages  21‐62  in  P.  S. Giller,  A. G. Hildrew,  and D. G.  Raffaelli,  editors.  Aquatic  ecology:  scale,  pattern  and  process.  Blackwell  Scientific.  Hubbell,  S.  P.  2001.  The  unified  neutral  theory  of  biodiversity  and  biogeography.  Princeton University Press.  Hurlbert, S. H. 1971. The non‐concept of  species diversity: a critique and alternative  parameters. Ecology 52:577‐586.  Huston, M. 1979. A general hypothesis of  species diversity. The American Naturalist  113:81‐100.  Huston,  M.  1994.  Biological  Diversity:  The  Coexistence  of  Species  on  Changing  Landscapes. Cambridge University Press, Cambridge, UK.  Kneitel, J. M., and J. M. Chase. 2004. Disturbance, predator, and resource interactions  alter container community composition. Ecology 85:2088‐2093.  Kondoh, M.  2001. Unifying  the  relationships  of  species  richness  to  productivity  and  disturbance.  Proceedings  of  the  Royal  Society  of  London  Series  B‐Biological  Sciences 268:269‐271.  Lake, P. S. 2000. Disturbance, patchiness, and diversity in streams. Journal of the North  American Benthological Society 19:573‐592.  Loreau, M., S. Naeem, P. Inchausti, J. Bengtsson, J. P. Grime, A. Hector, D. U. Hooper,  M. A. Huston, D. Raffaelli, B. Schmid, D. Tilman, and D. A. Wardle. 2001. Ecology  Chapter 3: Productivity, disturbance and diversity between landscapes      85    ‐  Biodiversity  and  ecosystem  functioning:  Current  knowledge  and  future  challenges. Science 294:804‐808.  Mackay, R. J. 1992. Colonization by Lotic Macroinvertebrates  ‐ a Review of Processes  and Patterns. Canadian Journal of Fisheries and Aquatic Sciences 49:617‐628.  Mackey,  R.  L.,  and  D.  J.  Currie.  2000.  A  re‐examination  of  the  expected  effects  of  disturbance on diversity. Oikos 88:483‐493.  Mackey,  R.  L.,  and  D.  J.  Currie.  2001.  The  diversity‐disturbance  relationship:  Is  it  generally strong and peaked? Ecology 82:3479‐3492.  McCabe, D. J., and N. J. Gotelli. 2000. Effects of disturbance frequency,  intensity, and  area on assemblages of stream macroinvertebrates. Oecologia 124:270‐279.  Mittelbach, G. G., C. F. Steiner, S. M. Scheiner, K. L. Gross, H. L. Reynolds, R. B. Waide,  M. R. Willig, S. I. Dodson, and L. Gough. 2001. What is the observed relationship  between species richness and productivity? Ecology 82:2381‐2396.  Morin, A., W. Lamourex, and  J. Busnarda. 1999. Empirical models predicting primary  productivity  from chlorophyll a and water temperature  for stream periphyton  and  lake  and  ocean  phytoplankton.  Journal  of  the  North  American  Benthological Society 18:299‐307.  Niyogi, D. K., M. Koren, C. J. Arbuckle, and C. R. Townsend. 2007. Stream communities  along  a  catchment  land‐use  gradient:  subsidy‐stress  responses  to  pastoral  development. Environmental Management 39:213‐225.  Petraitis, P. S., R. E. Latham, and R. A. Niesenbaum. 1989. The Maintenance of Species‐ Diversity by Disturbance. Quarterly Review of Biology 64:393‐418.  Pfankuch, D. J. 1975. Stream Reach  Inventory and Channel Stability Evaluation. USDA  Forest Service, Region 1, Missoula, Montana.  Quinn, J. M. 2000. Effects of pastoral development. Pages 208‐229  in K. J. Collier and  M.  J. Winterbourn,  editors.  New  Zealand  stream  invertebrates:  ecology  and  implications for management. The Caxton Press, Christchurch, New Zealand.  Resh, V. H., A. V. Brown, A. P. Covich, M. E. Gurtz, H. W. Li, G. W. Minshall, S. R. Reice,  A. L. Sheldon, J. B. Wallace, and R. C. Wissmar. 1988. The role of disturbance in  stream  ecology.  Journal  of  the North  American  Benthological  Society  7:433‐ 455.  Chapter 3: Productivity, disturbance and diversity between landscapes      86    Ricklefs, R.  E.  2004. A  comprehensive  framework  for  global patterns  in biodiversity.  Ecology Letters 7:1‐15.  Robinson, C. T., and G. W. Minshall. 1986. Effects of disturbance frequency on stream  benthic community structure in relation to canopy cover and season. Journal of  the North American Benthological Society 5:237‐248.  Rosenzweig, M.  L.  1995.  Species Diversity  in  Space  and  Time.  Cambridge University  Press, Cambridge, UK.  Rowe, L., B. Fahey, R. Jackson, and M. Duncan. 1997. Effects of land use on floods and  low flows. Pages 89‐102 in M. P. Mosely and C. P. Pearson, editors. Floods and  Droughts: the New Zealand Experience. New Zealand Hydrological Society Inc.,  Christchurch, New Zealand.  Roxburgh,  S.  H.,  K.  Shea,  and  J.  B.  Wilson.  2004.  The  intermediate  disturbance  hypothesis:  Patch  dynamics  and mechanisms  of  species  coexistence.  Ecology  85:359‐371.  Scholes, L., P. H. Warren, and A. P. Beckerman. 2005. The combined effects of energy  and  disturbance  on  species  richness  in  protist  microcosms.  Ecology  Letters  8:730‐738.  Steinman, A. D.,  and G. A.  Lamberti.  1996. Biomass  and  pigments  of  benthic  algae.  Pages 295‐314  in F. R. Hauer and G. A. Lamberti, editors. Methods  in Stream  Ecology. Academic Press, San Diego, CA.  Svensson,  J. R., M. Lindegarth, M. Siccha, M. Lenz, M. Molis, M. Wahl, and H. Pavia.  2007. Maximum  species  richness  at  intermediate  frequencies of disturbance:  Consistency among levels of productivity. Ecology 88:830‐838.  Tilman, D. 1999a. The ecological consequences of changes in biodiversity: A search for  general principles. Ecology 80:1455‐1474.  Tilman, D. 1999b. Global environmental  impacts of agricultural expansion: The need  for sustainable and efficient practices. Proceedings of the National Academy of  Sciences of the United States of America 96:5995‐6000.  Tonkin,  J. D., R. G. Death, and K.  J. Collier.  In prep. Do productivity and disturbance  interact to modulate macroinvertebrate diversity in streams?  Chapter 3: Productivity, disturbance and diversity between landscapes      87    Towns,  D.  R.,  and  W.  L.  Peters.  1996.  Leptophlebiidae  (Insecta:  Ephemeroptera).  Manaaki Whenua Press, Lincoln, New Zealand.  Townsend, C. R., M. R. Scarsbrook, and S. Doledec. 1997. The intermediate disturbance  hypothesis, refugia, and biodiversity  in streams. Limnology and Oceanography  42:938‐949.  Tscharntke,  T.,  A.  M.  Klein,  A.  Kruess,  I.  Steffan‐Dewenter,  and  C.  Thies.  2005.  Landscape  perspectives  on  agricultural  intensification  and  biodiversity  ‐  ecosystem service management. Ecology Letters 8:857‐874.  Waide, R. B., M. R. Willig, C.  F.  Steiner, G. Mittelbach,  L. Gough,  S.  I. Dodson, G. P.  Juday,  and  R.  Parmenter.  1999.  The  relationship  between  productivity  and  species richness. Annual Review of Ecology and Systematics 30:257‐300.  Walsh, C.  J., A. H. Roy,  J. W. Feminella, P. D. Cottingham, P. M. Groffman, and R. P.  Morgan. 2005. The urban stream syndrome: current knowledge and the search  for a cure. Journal of the North American Benthological Society 24:706‐723.  Williams, D. D.,  and H. B. N. Hynes.  1976.  The  recolonization mechanism of  stream  benthos. Oikos 27:265‐272.  Wilson,  S. D.,  and D.  Tilman.  2002. Quadratic  variation  in  old‐field  species  richness  along gradients of disturbance and nitrogen. Ecology 83:492‐504.  Winterbourn, M. J., and K. J. Collier. 1987. Distribution of benthic invertebrates in acid,  brown  water  streams  in  the  South  Island  of  New  Zealand.  Hydrobiologia  153:277‐286.  Winterbourn, M.  J., K.  L. D. Gregson, and C. H. Dolphin. 2000. Guide  to  the aquatic  insects of New Zealand. Entomological Society of New Zealand, Auckland.  Wolman, M.  J. 1954. A method of sampling coarse river bed material. Am. Geol. Un.  Trans. 35:951‐956.  Wootton,  J.  T.  1998.  Effects  of  disturbance  on  species  diversity:  A  multitrophic  perspective. American Naturalist 152:803‐825.      Chapter 4  Productivity‐diversity relationships in  streams and the role of periphyton  assemblages                      Chapter 4: Productivity‐diversity and periphyton assemblages      Abstract  89    There is much debate as to what the form of the productivity‐diversity relationship is,  particularly in lotic systems which have received few assessments of this relationship.   One factor that assessments of primary productivity do not account for  is the growth  form of algae that drives production.   Thus, we set out to (i) examine the productivity‐ diversity relationship in 24 streams of Cantabria, Spain, in July 2007 and (ii) determine  whether  this  relationship  was  underpinned,  and  better  explained,  by  specific  responses  to  the  form  of  the  periphyton  community.    Due  to  recent  evidence  in  streams suggesting productivity simply sets the upper limit to richness, rather than to  increase the effect of competitive  interactions, we hypothesized that diversity would  be a  log‐linear  function of productivity.   We predicted  that diversity would  respond  opposingly to two coarse measures of the periphytic community;  i.e., positively to %  diatom cover and negatively to % filamentous algae cover.  There was no relationship  between productivity and diversity in these streams but, as predicted, this relationship  was underpinned by responses to the growth form of periphyton community.  Diatom  cover was  the best predictor of  invertebrate diversity patterns.   Generally, diversity  responded positively to % diatom cover and negatively to % filamentous algae cover.   However,  results were  variable  and %  EPT  exhibited  a  greater  sensitivity  to  higher  levels  of  diatom  cover.    These  findings  highlight  two  important  implications:  firstly,  productivity‐diversity relationships in streams can be underpinned by interactions with  specific forms of periphyton.  Secondly, rapid assessment of periphytic forms is useful  for managers  as  these  coarse measures  are highly  relevant  to higher  trophic  levels.   We  suggest  a  threshold  of  40 %  filamentous  algae  cover  for managers wishing  to  minimise deleterious effects of eutrophication on stream communities.        Chapter 4: Productivity‐diversity and periphyton assemblages      Introduction  90    Although  the  relationship  between  productivity  and  diversity  is  a  central  theme  in  ecological  research  (Currie 1991, Abrams 1995, Waide et  al. 1999, Mittelbach et  al.  2001), we  are  far  from  reaching a  consensus on  the  form of  the  relationship either  empirically or  in  theory  (Waide et al. 1999, Mittelbach et al. 2001).   Unimodal  (e.g.  Grime 1973, Huston 1979, Rosenzweig 1992, 1995, Leibold 1999, Waide et al. 1999,  Mittelbach et al. 2001) and  linear  (e.g. Currie and Paquin 1987, Currie 1991, Abrams  1995, Waide et al. 1999, Gaston 2000, Mittelbach et al. 2001)  increases  in diversity  with  increasing productivity prevail  as  the most  common  forms of  this  relationship.   Differences    in the observed patterns may be a result of several factors  including the  spatial  scale of observation  (Currie  1991, Mittelbach  et  al.  2001, Chase  and  Leibold  2002), disturbance (Huston 1979, 1994), and history of community assembly (Fukami  and Morin 2003).   However,  the relationship has also been shown  to differ between  ecosystems and organisms studied (Waide et al. 1999, Mittelbach et al. 2001).   These  differences  suggest  individual  patterns  are  driven  by  underlying  ecosystem  specific  interactions.    Furthermore,  the  methods  used  to  collate  the  information  on  the  productivity‐diversity  relationship  have  recently  come  under  heavy  criticism  due  to  inconsistent classification and errors in meta‐analyses (Hillebrand and Cardinale 2010,  Whittaker 2010).    While abiotic forces (floods, hurricanes etc.) are considered to dominate biotic  forces (competition, predation) in many systems, plant‐herbivore relationships remain  the  basis  of most  ecosystems.    As  such  the  control  of  diversity  is  likely  to  be  an  interaction between disturbance and the productivity of an ecosystem (Huston 1979,  1994,  Kondoh  2001).   No  system  is more  heavily  influenced  by  regular  disturbance  events than streams (e.g. flooding) (Resh et al. 1988, Lake 2000, Death 2008).   Death  (2002) suggests that floods remove taxa and their resources and opens the habitat for  recolonisation whilst primary productivity  sets  the upper  limit  to how many animals  can return to these habitats after the disturbance.  Recent support has been found for  Death’s  hypothesis  but  these  studies  found  no  evidence  of  interactive  effects  of  disturbance and productivity  (Tonkin and Death  In prep., Tonkin et al.  In prep.), but  Chapter 4: Productivity‐diversity and periphyton assemblages      91    rather that diversity is a result of the additive effects of disturbance and productivity.   Without  the  interactive  effects  of  disturbance,  making  sense  of  the  productivity‐ diversity  relationship  in  streams  should  be  clear‐cut.   However,  compared  to  lentic  systems and  in fact most other environments, few studies have  investigated whether  higher  productivity  leads  to  greater  diversity  in  lotic  systems.    The  few  to  look  specifically at this have found both unimodal (Death and Zimmermann 2005) and log‐ linear  (Death 2002, Barquin 2004, Tonkin and Death  In prep., Tonkin et al.  In prep.)  increases in diversity with productivity.    Assessing  the  relationship  between  primary  productivity  and  diversity  is  essentially  testing  the  response  of  higher  trophic  levels  to  the  rate  of  energy  production  at  lower  trophic  levels.    Primary  productivity  is  typically  provided  in  streams by periphytic algae.   Algal  forms vary greatly  in  streams  from prostrate and  stalked diatoms through to filamentous green algae (Allan 1995), all of which respond  differently  to  environmental  conditions.    It  is  now  common  practice  to  use  the  periphyton community as an  index for biomonitoring environmental conditions (Kelly  and Whitton 1995, Pan et al. 1996, Kelly 1998b, a, Kelly and Whitton 1998, Hill et al.  2000).   Typically the focus of these assessments of biotic  integrity have been diatoms  (Kelly and Whitton 1995, Pan et al. 1996, Kelly 1998b), but Whitton and Kelly  (1995)  advocated the use of the full community of plants including bryophytes.      Not  only  do  various  forms  of  periphytic  algae  respond  differently  to  environmental  conditions  but  they  can  provide  diverse  habitat  and  resources  for  higher  trophic  levels.    Different  periphyton  growth  forms  can  also  fulfil  different  functional  roles  in  benthic  communities  (Steinman  et  al.  1992).    When  levels  of  periphyton  reach greater densities and epilithic diatoms are  replaced by macroalgae  such as filamentous green algae, interactions between grazers and periphyton can shift  from  simple plant‐herbivore  interactions  to more  complex  relationships.   As well  as  providing  food  for  a  few  specialist  taxa, macroalgae  can  both  provide  and  remove  habitat  and  compete  for  space  with  invertebrates.    Dudley  et  al.  (1986)  classed  invertebrates  into  those  negatively  affected  by macroalgae  due  to  competition  for  space,  positively  affected  due  to  habitat  provision,  and  positively  affected  by  food  provision.   This can be reflected  in the typical shift from pollution (nutrient) sensitive  Chapter 4: Productivity‐diversity and periphyton assemblages      92    taxa associated with thin diatom films, to pollution tolerant taxa and filamentous algae  growth  forms often  associated with nutrient eutrophication  (Suren et  al. 2003).   Of  course, grazing can have a large impact on periphyton in aquatic systems and this top‐ down  control  has  been  the  central  focus  of  periphyton‐invertebrate  community  relationships to date (Hillebrand 2009).  This grazing control can differ between growth  forms of periphyton (Feminella and Hawkins 1995), such as that between diatom and  filamentous  forms  (Suren  and  Riis  2010).    Lamberti  (1989)  suggests  this  differential  response of algae may  initially occur where productive capacity of algae  is high, but  may be later regulated through the arrival of specialist grazers, or physical disturbance  (Fisher et al. 1982).  This indicates that vigilance is needed when inferring top‐down or  bottom‐up  control  as  it  is  likely  to  change  through  time  and may be dependent on  animals present in the system at hand.    Compared  to  other  ecosystems,  the  productivity‐diversity  relationship  has  received relatively little attention from lotic ecologists.  Thus, we set out to (i) test the  response  of  stream  invertebrate  diversity  to  primary  productivity;  and  (ii)  because  productivity measurements do not account for variation in the form of producers, we  examine  if  this  productivity‐diversity  relationship  can  be  better  explained  by  underlying  responses  to  different  forms  of  periphyton  (i.e.  diatoms/periphytic mats  and filamentous green algae).  We also use a common stream specific metric, % EPT, to  assess whether  this metric  is more  sensitive  to  environmental  gradients  in  streams  than simple diversity measures.  As a result of previous work in streams (Death 2002,  Barquin  2004,  Tonkin  and Death  In  prep.,  Tonkin  et  al.  In  prep.),  and  the  fact  that  productivity does not appear to increase rates of competitive exclusion in streams as it  often  does  in  lakes, we  hypothesise  that  diversity will  increase  logarithmically with  increasing  productivity.    However,  we  predict  that  this  relationship  will  be  underpinned by particular responses to different forms of periphyton.  Specifically, due  to  the view  that diatoms are considered  favourable  to grazers and  filamentous algal  forms poor habitat for many invertebrates (Suren and Riis 2010), diversity will respond  positively to % diatom cover and negatively to % filamentous algae cover.      Chapter 4: Productivity‐diversity and periphyton assemblages      Methods  93    Study sites  Twenty  four  streams were  sampled  in  the Cantabrian Mountains of Northern  Spain  (Fig. 1, Appendix 1). The Cantabrian Mountains span ~483 km east to west along the  northern  coast of  Spain  reaching 2,648 m  asl  at Torre de Cerredo.   Average  rainfall  ranges from ~1,200 to 1,600 mm p.a. depending on location within the region.  Land‐ use  surrounding  sampling  sites  varied  from  Atlantic  deciduous  forest  consisting  predominantly of oak (Quercus spp.) and European beech (Fagus spp.) to pasture and  small urban  settlements.   Sites were  selected  from  six  river catchments: Rio Besaya,  Rio Saja, Rio Pas, Rio Pisueña, Rio Nansa, and Rio Ebro.   Within each catchment sites  were  selected  a  priori  so  that  one  low  and  one  high  productivity  site  in  close  geographic  proximity  were  sampled.    As  these  were  selected  prior  to  sampling,  productivity estimates  for  the selection of a priori high and  low productivity streams  were based on visual estimates of periphyton which are detailed below.  All sites were  cobble  bottom  streams.    Altitude  of  the  sites  ranged  from  163  to  1061 m  asl  and  average channel width ranged from 1.9 to 30.7 m (Appendix 1).        Physicochemical variables  Depth  and water  velocity were  recorded with  a Marsh‐McBirney  flowmate  current  meter  at  five  equidistant  points  along  the  thalweg.  Conductivity,  temperature,  dissolved oxygen concentration and percent saturation, and pH were measured using a  YSI  556  MPS  meter.    Nitrate,  phosphate  and  ammonia  were  calculated  using  a  Beckman Coulter DU® Series 700 UV/Vis Scanning Spectrophotometer.   Substrate size  composition  was  assessed  by  sampling  100  stones  using  the  Wolman  walk  methodology (Wolman 1954). This was converted to a substrate size    index following  Jowett  and  Richardson  (1990).    Substrate  heterogeneity  was  assessed  using  the  Shannon diversity  index.   Bed stability was assessed using  the bottom component of  the  Pfankuch  stability  index  (Pfankuch  1975).    Flow  type  of  each  site was  assessed  visually as percentage of still, backwater, pool, run or riffle over a 100 m reach.  Leaf  litter percentage, riparian vegetation and % canopy cover were also assessed visually.   Chapter 4: Productivity‐diversity and peri   phyton assemblages        Figure 1: M Latitudes a   Biologic Periphyton Primary  p biomass ( five stone cool  and  using 90 % Coulter DU concentra estimated of  the  sto productivi related (M ap showing nd longitude al collec  and prima roductivity Morin et al s (mean ar dark  befor  acetone a ® Series 70 tion  follow  following G ne  availab ty and chlo orin et al.   the  locatio s are given  tions  ry product   was  not  . 1999).   Bi ea: 60 cm2 e  being  fro t 5°C for 24 0 UV/Vis S ing  Steinm raham et a le  for  per rophyll a  f 1999, Ton 9 n of 24 stre in Appendix  ivity  measured omass was ) collected  zen.    Chlo  h in the d canning Sp an  and  L l. (1988) a iphyton  gr rom stone kin and De 4  ams  in Can 1.      directly  b  measured  randomly f rophyll  a  a ark.  Absor ectrophoto amberti  (1 nd then ha owth.    The s  in stream ath  In prep tabria, Spain ut  estima by extracti rom each s nd  phaeop bances we meter and  996).    Sto lved to corr   relations s has been .).   Morin e   , sampled  in  July 2007.   ted  from  ng chlorop ite.   These hytin were re read on  converted  ne  surface ect for the  hip  betwee   found  to  t al. (1999 periphyton hyll a from  were kept   extracted a Beckman to pigment   area  was proportion n  primary be strongly ) reviewed                       Chapter 4: Productivity‐diversity and periphyton assemblages      95    this  relationship  and  found  chlorophyll  a  extracted  from  periphyton  and  primary  productivity were strongly  log  linearly related (r2 = 0.63).   Tonkin et al. (In prep.) also  found  a  strong  linear  link between  chlorophyll a biomass  accumulation on  tiles  and  chlorophyll a content of stones r2 = 0.74).   We calculated two other metrics to assess  periphyton  communities:  percent  cover  of  diatoms  and mats,  and  percent  cover  of  filamentous  algae.    These  were  assessed  visually  along  three  transects  across  the  stream bed using  rapid assessment protocols  (Biggs and Kilroy 2000).   Subsequently,  we grouped all algal forms into two groups: periphytic films and mats and filamentous  forms.   For  the  sake of  simplicity, periphytic mats will be  called diatoms although  it  includes all crustose, prostrate and stalked algae forms.        Macroinvertebrates  Five 0.1 m2 500 µm mesh Surber samples were collected at random from riffles at each  site.    In  the  field  samples were  preserved  in  10 %  formalin.    In  the  laboratory  the  samples  were  washed  through  500  µm  and  1  mm  Endecott  sieves  before  being  identified and enumerated to the lowest possible taxonomic level.  Invertebrates were  mostly  identified  to morphospecies,  however, where  possible morphospecies were  identified using available keys (e.g. Tachet et al. 2000).    The number of animals per 0.1 m2 (density) was calculated for each  individual  sample and averaged per site as was the number of taxa (richness).  Rarefied richness  (ES[N]) was calculated in Primer v6 (Clarke and Gorley 2006) for 261 individuals which  was  the  lowest  average  number  of  animals  at  a  site.    Rarefaction  accounts  for  the  passive increase in the number of taxa collected with increasing number of individuals  collected  (Sanders  1968,  Hurlbert  1971).    This,  in  effect,  standardises  sites  by  predicting  richness  per  a  set  number  of  animals  rather  than  a  set  area.    The  final  community  metric  used  was  the  percent  of  Ephemeroptera,  Plecoptera  and  Trichoptera (EPT) animals per sample (Lenat 1988).       Statistical methods  To  test  whether  diversity,  periphyton  metrics  or  other  physicochemical  variables  differed between high and  low productivity  sites, Analysis of Variance  (ANOVA) was  Chapter 4: Productivity‐diversity and periphyton assemblages      96    used in Statistix (Statistix 8 ©, Analytical Software, Tallahassee, FL, U.S.A.).  We used a  randomised complete block  (RCB) design using productivity groups as  the  treatment  and catchments as blocks.   This method  reduces experimental error by dividing  into  similar  groups  (i.e.  catchment).   Analysis of  Similarities  (ANOSIM)  (Clarke 1993) was  performed  on  normalised  physicochemical  data  using  Euclidean  distances  to  assess  whether site characteristics differed between high and low productivity treatments in  Primer v6 (Clarke and Gorley 2006).    To visualise differences  in community structure between the two productivity  treatments  and  catchments, we  carried  out  a  non‐metric multi  dimensional  scaling  (NMDS)  ordination  on  log  (x  +  1)  data  using  Bray‐Curtis  similarity  and  tested  for  significance using analysis of similarities  (ANOSIM)  (Clarke 1993)  in Primer v6  (Clarke  and  Gorley  2006).    Using  log  (x+1)  transformed  data,  we  used  the  similarity  percentages (SIMPER) (Clarke 1993) procedure  in Primer v6 (Clarke and Gorley 2006)  to  assess which  taxa were  contributing most  to  differences  in  community  structure  between  high  and  low  productivity  sites.    Regression  was  also  used  to  examine  relationships between primary productivity, percent diatom and filamentous cover and  invertebrate metrics, and some physicochemical variables using Statistix (Statistix 8 ©,  Analytical  Software, Tallahassee,  FL, U.S.A.).   Where  required, data were  log  (x + 1)  transformed to remove heteroscedasticity.    To analyse which variables best predicted invertebrate and periphyton metrics  we used M5 pruned regression trees using Weka 3.6 (Hall et al. 2009).  These were re‐ run using 5‐fold cross validation to test whether the model was applicable outside the  particular sites used.  Classification and Regression Tree (CART) analysis was also run to  easures of periphyton from the remaining physiochemical variables.   predict the four m   Results  Physicochemical characteristics  Conductivity ranged from 68 to 402 µS cm‐1 (Appendix 1).  Conductivity and dissolved  oxygen  were  the  only  physicochemical  variables  to  differ  between  high  and  low  productivity sites (Table 1).   Conductivity was nearly twice as high on average at high  Chapter 4: Productivity‐diversity and periphyton assemblages      97    productivity sites  (low mean = 157 µS cm‐1, high mean = 255 µS cm‐1) and dissolved  oxygen averaged 9.85 mg l‐1 at low productivity sites compared to 10.44 mg l‐1 at high  productivity  sites  (Table  1;  Appendix  1).    The  bottom  component  of  the  Pfankuch  stability index did not differ between low and high productivity sites (Table 1).  There  was  no  difference  in  physicochemical  data  between  productivity  treatments  with  either all variables included (R = 0.16, P = 0.34) or on the remainder of variables with  conductivity and dissolved oxygen removed (R = ‐0.05, P = 0.85).      Table 1: Randomised  complete block analysis of variance  (ANOVA)  results  for differences  in  mean physicochemical and biological variables between a priori selected productivity groups in  24 streams of Cantabria, Spain, July 2007.  Productivity group is the treatment and catchment  the block.  D.O. = dissolved oxygen, O.H. Cover = overhead cover.   Degrees of freedom for the  F‐test for all variable is 1, 17.             Productivity     Low (S.E.)  High (S.E.)  F  P  Pfankuch bottom  31.92 (2.04)  33.58 (2.88)  0.21  0.65  Altitude  393.17 (86.21)  433.17 (77.49)  0.72  0.41  pH  8.81 (0.06)  8.91 (0.07)  1.49  0.24  Conductivity (µs cm‐1)  156.67 (20.75)  255.25 (24.95)  10.92  0.004  Temperature (°C)  15.62 (0.57)  14.69 (0.52)  1.38  0.26  D.O. (mg l‐1)  9.85 (0.12)  10.44 (0.19)  7.58  0.01  D.O. (%)  98.93 (0.89)  102.03 (2.11)  1.63  0.22  O.H. Cover (%)  35.42 (8.52)  29 (6.93)  0.31  0.59  Velocity (m s‐1)  0.4 (0.04)  0.39 (0.04)  0.07  0.79  Depth (cm)  19.7 (1.24)  18.02 (1.54)  1.02  0.33  Width (m)  7.39 (1.08)  8.05 (2.29)  0.18  0.68  Substrate size index  147.04 (6.8)  131.17 (9.85)  1.94  0.18  Substrate heterogeneity  1.97 (0.05)  1.99 (0.03)  0.2  0.66  Slope (m 100m‐1)  5.08 (0.94)  5.28 (0.83)  0.02  0.89  Nitrate (mg l‐1)  0.69 (0.18)  0.93 (0.14)  2.07  0.17  Phosphate (mg l‐1)  0.43 (0.08)  0.37 (0.04)  0.32  0.58  Ammonia (mg l‐1)  0.05 (0.01)  0.07 (0.02)  0.87  0.36  Chlorophyll a (µg cm‐2)  4.47 (0.84)  7.22 (1.05)  3.91  0.07  Diatom (%)  57.22 (7.71)  39.31 (7.49)  4.97  0.04  Filamentous algae (%)  3.56 (1.48)  44.5 (8.77)  18.9  0.0004  Total peri. cover (%)  60.78 (8.38)  83.81 (6.97)  4.9  0.04  Bryophyte (%)  5.5 (4.96)  1.39 (0.96)  0.55  0.47  No. of taxa  16.15 (0.77)  15.05 (0.91)  1.24  0.28  Log (no. of animals)  2.54 (0.05)  2.76 (0.1)  3.25  0.09  ES(261)  15.37 (0.69)  13.48 (0.88)  3.43  0.08  % EPT animals  0.63 (0.05)  0.5 (0.05)  4.19  0.06      Chapter 4: Productivity‐diversity and periphyton assemblages      98    Periphyton   Primary productivity, assessed as mean  chlorophyll a,  ranged  from 2.58  to 15.35 µg  cm‐2 (Appendix 1).  Mean chlorophyll a was greater at high productivity sites than low  productivity  sites but  the difference was only  significant at  the 10 %  level  (Table 1).   Percent diatom cover was lower and filamentous algae cover was more than ten times  greater at high productivity sites. Chlorophyll a was negatively correlated with percent  diatom cover (r = ‐0.41, P = 0.045) and positively correlated with filamentous cover (r =  0.63, P = 0.001), and filamentous cover was negatively associated with diatom cover (r  = ‐0.5, P = 0.013).  Chlorophyll a (F1, 22 = 15.64, P = 0.0007, r 2 = 0.416, y = 0.81 + 0.02x)  and percent filamentous algae cover (F1, 22 = 29.05, P <0.0001, r 2 = 0.569, y = ‐26.05 +  0.24x)  increased  linearly  whereas  percent  diatom  cover  declined  with  increasing  conductivity (F1, 22 = 5.99, P = 0.02, r 2 = 0.214, y = 76.37 ‐ 0.14x; Fig. 2).       Community composition  Ordination  and  ANOSIM  on  log  (x  +  1)  transformed  invertebrate  data  indicated  community composition did not differ between high and  low productivity treatments  (R = 0.05, P = 0.16) or between catchments (R = 0.002, P = 0.47; Fig. 3).   Filamentous  algae cover and chlorophyll a were positively correlated with MDS 1 (r = 0.47 and 0.23  respectively) and negatively correlated with MDS 2  (r =  ‐0.19 and  ‐0.28 respectively),  whereas diatom cover was negatively correlated with MDS 1 (r = ‐0.27) and positively  correlated with MDS 2 (r = 0.52).        Chapter 4: Productivity‐diversity and periphyton assemblages      99      Figure 2: Periphyton metrics as a  function of conductivity  in 24  streams  in Cantabria, Spain,  July 2007.  (a) Chlorophyll a, (b) diatom cover, (c) filamentous algae cover.        Chapter 4: Productivity‐diversity and peri   phyton assemblages      Figure 3: N invertebrat triangles a above sym and Ebro).  text.  2D st   Te compositi (Table 2).  than  25  However,  1.36, P = 0 0.17; Perc low produ levels of d increase w Baetis spp 0.37).    T difference on‐metric m e communit re a priori se bols represe  Results of  ress = 0.19.   n  taxa  con on  (log  [x+  The simuli percent  to both densi .26; Perce ent: F1, 22 = ctivity  site iatom  cov ith increas . peaked a he  amphip  in inverteb ultidimens y data colle lected  low  nt the six ca ANOSIM tes   tributed  8 1]  transfor id Prosimul   the  diffe ty and per nt: F1, 22 = 3  0.22, P = 0 s  (Fig. 4).    er  (F2, 21 =  ing filamen t intermedi od  Echino rate comm 1 ional scaling cted from 2 productivity tchments sit ting betwee 6  %  to  t med)  betw ium spp. an rence  in  p cent compo .28, P = 0.0 .64) were  Percent Pr 7.39, P = 0 tous cover ate levels o gammarus  unities bet 00    (NMDS) or 4 streams in  sites and c es belong to n groups an he  differen een  high  a d mayfly B roductivity sition of P 8) and Bae not signific osimulium  .004,  r2 =   (F2, 21 = 2.9 f diatom c spp.  cont ween prod dination on   Cantabria,  losed are hig  (Besaya, Sa d axis corre ce  in  inv nd  low  pr aetis spp. e   treatmen rosimulium tis spp. (De antly differ spp. was  lo 0.41)  and  4, P = 0.07 over (F2, 21  ributed  ov uctivity tre log  (x + 1)  t Spain, July 2 h productiv ja, Pas, Pisu lations are  ertebrate  c oductivity  ach contrib ts  (Fig.  4;  spp. (Dens nsity: F1, 22 ent betwee west at  in exhibited  a 5, r  = 0.22 2 = 6.25, P = er  9  perce atments (F ransformed 007.  Open ity.   Names eña, Nansa, given  in the ommunity treatments uted more   Table  2). ity: F1, 22 =  = 2.04, P = n high and termediate  quadratic ).  Percent  0.007, r2 = nt  to  the ig. 4; Table                                        Chapter 4: Productivity‐diversity and periphyton assemblages      101    2).   However, neither density  (F1, 22 = 2.32, P = 0.14) or  relative abundance  (F1, 22 =  0.76,  P  =  0.39)  were  significantly  different  between  productivity  treatments.   Moreover,  percent  Echinogammarus  spp.  did  not  respond  to  either  diatom  (F1, 22  =  1.49, P = 0.24, r2 = 0.06) or filamentous cover (F1, 22 = 1.98, P = 0.17, r 2 = 0.08).      Table 2: Differences in community composition between the a priori productivity groups in 24  streams sampled in Cantabria, Spain, July 2007.  Table shows Similarity percentages (SIMPER)  showing  the  ten  taxa  contributing  to  the  greatest  difference  between  low  and  high  productivity streams.  Av.Abund = average abundance, Av.Diss = average dissimilarity, Diss/SD  =  standard  deviation  of  dissimilarity,  Contrib%  =  %  contribution,  Cum.%  =  cumulative  %  contribution.                         Low P  High P                                     Species  Av.Abund  Av.Abund  Av.Diss  Diss/SD  Contrib%  Cum.%  Prosimulium spp.  50.53  699.78  15.32  0.85  28.11  28.11  Baetis spp.  196.53  365.7  14.01  1.58  25.69  53.8  Echinogammarus spp.  20.77  56.87  5.28  0.85  9.69  63.49  Orthocladiinae b  17.8  119.9  3.13  0.92  5.75  69.23  Ephemerella spp.  22  18.43  2.1  0.93  3.86  73.09  Bythinella spp.  12.52  8.45  1.84  0.66  3.38  76.47  Protonemura spp.  19.87  8.92  1.56  1.09  2.86  79.34  Hydropsyche spp.  20.87  5.53  1.51  0.81  2.78  82.11  Orthocladiinae a  1.82  41.2  0.94  0.7  1.72  83.83  Oligochaeta  4.23  21.75  0.93  0.84  1.71  85.54      Chapter 4: Productivity‐diversity and periphyton assemblages      102      Figure 4: (a) Relative abundance and (b) mean abundance of all taxa between a priori selected  high  and  low  productivity  sites  in  24  streams  in  Cantabria,  Spain,  July  2007.    (c)  Relative  abundance  and  (b)  mean  abundance  of  the  three  most  dominant  taxa  between  a  priori  selected high and low productivity sites.      Density and diversity patterns  Rarefied richness and % EPT animals were higher and total number of animals lower at  low  productivity  sites,  although  these  differences were  only  significant  at  the  10 %  level (Table 1).   The number of taxa did not differ between high and  low productivity  sites.  The only aspect of the invertebrate community to respond to chlorophyll a was  the number of individuals which increased monotonically with increasing biomass (Fig.  5; Table 3).   Taxonomic  richness,  rarefied  richness and  the % EPT animals were not  related to chlorophyll a (Fig. 5; Table 3).      Chapter 4: Productivity‐diversity and periphyton assemblages      103      Figure 5: (a) Mean (± 1 S.E.) taxonomic richness, (b) number of animals, (c) rarefied richness,  and  (d) percent EPT animals as a function of chlorophyll a  in 24 streams of Cantabria, Spain,  July 2007.      Density  and  diversity  measures  exhibited  opposing  responses  to  the  two  growth forms of periphyton measured.  In general, the response to increasing diatom  cover was mainly  positive whereas, with  the  exception  of  density,  the  response  to  increasing filamentous cover was negative.   Taxonomic richness and rarefied richness  increased log‐linearly with increasing substrate cover of diatoms (Fig. 6; Table 3).  The  number of individuals was not related to percent diatom cover.  The percentage of EPT  animals appeared more sensitive  to higher percentages of diatom cover and peaked  strongly  at  intermediate  levels  and  declined  at  higher  levels  of  diatom  cover.   Taxonomic  richness  was  not  related  to  percent  filamentous  algae  cover,  however  density of  individuals exhibited quadratic  increase with  increasing  filamentous cover.   Chapter 4: Productivity‐diversity and periphyton assemblages      104    Both  rarefied  richness and % EPT animals  responded negatively  to  filamentous algae  exhibiting a curvilinear decline with increasing filamentous cover.        Table3:  Results  for  regression  analysis  for  taxonomic  richness,  number  of  animals,  rarefied  richness,  and  percent  EPT  animals  against  periphyton metrics  for  24  streams  in  Cantabria,  Spain, July 2007.   a) Chlorophyll a, b) percent diatom cover, and c) percent filamentous algae  cover.  Degrees of freedom for linear and log‐linear models are 1, 22 and for quadratic models  are 2, 21.  AIC = Akaike’s information criterion for the selection of the best model between log‐ linear and quadratic curves.  Lowest values represent the best model.                      F (AIC)  P  r2 Equation  a) Chlorophyll a (µg cm‐2)  No. of taxa  0.05  0.83  0.002  Non‐significant  Log (no. of animals)  5.31  0.03  0.194  y = 2.43 + 0.037x  ES(261)  0.91  0.35  0.040  Non‐significant  % EPT animals  1.14  0.30  0.049  Non‐significant  b) Diatom cover (%)  No. of taxa  9.62 (42.8)  0.01  0.304  y = 7.55 + 2.2 ln(x)      Quadratic  4.91 (43.1)  0.02  0.319  y = 10.7 + 0.19x ‐ 0.002x2 Log (no. of animals)  0.62  0.44  0.027  Non‐significant  ES(261)  12.77 (42.3)  0.002  0.367  y = 5.73 + 2.37 ln(x)      Quadratic  6.38 (42.8)  0.007  0.378  y = 9.08 + 0.21x ‐ 0.0016x2 % EPT animals  9.36  0.001  0.471  y = 0.15 + 0.02x ‐ 0.0002x2 c) Filamentous algae cover (%)  No. of taxa  1.59  0.22  0.067  Non‐significant  Log (no. of animals)  13.71  0.0002  0.566  y = 2.59 ‐ 0.011x + 0.0002x2 ES(261)  4.22  0.03  0.287  y = 14.93 + 0.059x ‐ 0.001x2 % EPT animals  4.19  0.03  0.285  y = 0.58 + 0.006x ‐ 0.0001x2     Chapter 4: Productivity‐diversity and periphyton assemblages      105      Figure 6:  (a, b) Mean  (± 1 S.E.)  taxonomic  richness,  (c, d) number of animals,  (e,  f)  rarefied  richness, and (g, h) percent EPT animals as a function of percent diatom cover (a, c, e, g), and  percent filamentous algae cover (b, d, f, h) in 24 streams of Cantabria, Spain, July 2007.        Chapter 4: Productivity‐diversity and periphyton assemblages      106    Classification and regression trees  Regression tress predicted taxa number well (82 % variation explained  in full data set  and 51 %  in cross validation, Table 4).   Sites were split  into four groups at two  levels  (Fig. 7; Table 4).  As predicted, the type of algae was important in predicting diversity  with the initial split grouping sites either side of 42 % diatom cover.  High diatom sites  (17.2  taxa)  had  greater  diversity  than  low  diatom  sites  (13.7  taxa)  (F1, 22 =  4.3,  P  =  0.001).   The sites with  less than 42 % diatom cover (LM1 and LM2) were further split  into stable or unstable groups based on the Pfankuch bottom component score of 25  (Fig.  7;  Table  4).    The  low  diatom  sites  were  explained  by  linear  model  1  and  2  consisting of the Pfankuch bottom component, percent diatom, altitude, and substrate  size (Fig. 7; Table 4).  The eight unstable sites (LM2) had the lowest taxonomic richness.   The high diatom sites (LM3 and LM4) were further split based on phosphate levels and  were predicted by chlorophyll a, diatom cover and altitude.  The high phosphate group  (LM4), which included 7 sites, had 15.9 taxa per site.    The  remaining  measures  of  density  and  diversity  related  to  measured  environmental variables in a more straightforward single linear model.  The number of  animals was best predicted by  altitude  and  conductivity  (Table 4). Rarefied  richness  was predicted by diatom cover (53 % variation explained  in full set and 29 %  in cross  validation;  Table  4).   None  of  the measured  variables  could  predict %  EPT  animals  (Table 4).    Conductivity was  the  best  predictor  of  primary  productivity  and  periphyton  cover and was included in the prediction models of all three periphyton metrics (Table  4).  Division of sites was not required for the prediction of these metrics.  Chlorophyll a  was predicted well by pH, conductivity and velocity (83 % variation explained in full set  and 56 % in cross validation; Table 4).  Diatom cover was predicted by conductivity and  temperature (62 % variation explained in full set and 29 % in cross validation; Table 4).   Conductivity predicted  filamentous algae cover percentage  (75 % variation explained  in full set and 60 % in cross validation; Table 4).        Chapter 4: Productivity‐diversity and periphyton assemblages      107    Table  4:  Fit  statistics  showing  the  results  of M5  pruned  classification  and  regression  tree  (CART) analysis predicting invertebrate community metrics and four measures of periphyton in  24 streams in Cantabria, Spain, July 2007.  S = number of taxa, LogN = log (number of animals),  ES(261) = rarefied richness for 261  individuals, LM =  linear model, RMSE = root mean square  error, RRSE = root relative square error.                         (a)  Training data     5‐fold cross validation        Correlation  coefficient  RMSE  RRSE  Correlation coefficient  RMSE  RRSE  S  0.815  1.71  60.3  0.51  2.45  85.13  LogN  0.6  0.23  80  0.078  0.34  112.78  ES(261)  0.529  2.37  84.84  0.289  2.85  100.64  % EPT animals  0  0.17  100  ‐0.319  0.27  147.75  Chlorophyll a  0.83  1.92  55.85  0.559  2.96  84.78  Diatom  0.615  21.08  78.83  0.294  27.76  99.55  Filamentous  0.754  19.19  65.64  0.602  24.46  81.94  (b)  Model  Equation  S  LM1  S = ‐0.07(Pfankuch) + 0.03(Diatom) + 0.002 (Alt.) + 4.57(Substrate) + 5.63  LM2  S = ‐0.06(Pfankuch) + 0.03(Diatom) + 0.0008(Alt.) + 5.42(Substrate) + 3.42  LM3  S = 0.17(Chlorophyll a) + 0.03(Diatom) + 0.003(Alt.) + 12.94  LM4  S = 0.17(Chlorophyll a) + 0.03(Diatom) + 0.003(Alt.) + 12.60  LogN  LM1  LogN = 0.0004(Alt.) + 0.002(Cond.) + 2.112  Rarefaction  LM1  ES261 = 0.06(Diatom) + 11.75  % EPT animals  LM1  % EPT animals = 0.57  Chlorophyll a  LM1  Chlorophyll a = 5.82(pH) + 0.02(Cond.) ‐ 11.04(Velocity) ‐ 45.14  Diatom  LM1  Diatom = ‐0.11(Cond.) + 5.98(Temp.) ‐ 20.17  Filamentous  LM1  Filamentous = 0.24(Cond.) ‐ 26.05       Figure 7: M5 pruned regression tree generated for the prediction of the number of taxa in 24  streams  of  Cantabria,  Spain,  July  2007.    The  tree  shows  the  relationship  between  three  explanatory variables and the levels at which the splits occur.  Mean (± 1 S.E.) number of taxa  per  site  are  given  in each  terminal node with  the number of  sites predicted  for each node  given  in  parentheses.    Diatom  =  %  diatom  cover,  Pfankuch  =  bottom  component  of  the  rogen = mg LPfankuch stability index, Nit fit statistics and 4b for equat   Discussion  ‐1.  LM1 – LM4 = linear model 1 ‐ 4.  See Table 4a for  ions for each linear model.    There was no  relationship between periphyton biomass and  invertebrate diversity  in  this study.  This suggests there was no clear relationship between primary productivity  and  invertebrate diversity as primary productivity and benthic periphyton biomass  is  strongly  linked  in  streams  (Morin  et  al.  1999,  Tonkin  and  Death  In  prep.).   Meta  analyses  have  shown  that  approximately  one  third  of  all  studies  looking  at  the  productivity‐diversity relationship find no significant link between the two, but this can  vary with spatial scale, community type and between plants and animals (Waide et al.  1999,  Mittelbach  et  al.  2001).    It  appears  that  this  lack  of  relationship  between  productivity and diversity  is more common at smaller scales, such as this study, with  links between the two more commonly found at larger spatial scales (Mittelbach et al.  2001).   We must interpret this data with caution as there is now considerable debate  about  the  robustness  of  the  meta‐analyses  used  to  collate  data  addressing  this  question (Hillebrand and Cardinale 2010, Whittaker 2010).  Recently, several studies in  stream  communities  have  found  log‐linear  increases  in  diversity  with  productivity  (Death 2002, Barquin 2004, Tonkin and Death  In prep., Tonkin et al.  In prep.).   This,  Chapter 4: Productivity‐diversity and periphyton assemblages      109    and the belief that increased productivity does not lead to greater rates of competitive  exclusion  in  streams,  led  us  to  predict  this  log‐linear  trend  would  occur  in  these  Spanish  streams;  however  this  was  not  the  case.    Lamberti  et  al.  (1989)  suggest  changes  in  herbivore  biomass  may  better  represent  the  production  of  stream  communities  (i.e. primary productivity)  than algal biomass due  to  the  strong  role of  herbivory  suppressing  algal  biomass  in  streams,  although  this  does  not  account  for  disturbance.      Although  invertebrate  communities  did  not  respond  clearly  to  changes  in  primary  productivity,  the  form  of  the  periphyton  community  was  important  in  determining diversity patterns.   There has been extensive  research on  the effects of  grazers on algal communities, but other  than grazer  specific  responses  (e.g. Gresens  and  Lowe  1994, Maasri  et  al.  2008),  the  reverse  effects  of  algal  assemblages  on  invertebrate  communities  has  received  little  direct  attention  in  streams  (but  see  Dudley et al. 1986, Koksvik and Reinertsen 2008).  However, it is clear that the type of  periphyton  has  strong  influences  on  the  structure  of  stream  benthic  communities  (Dudley et al. 1986, Koksvik and Reinertsen 2008), and grazing communities can in fact  grow at different  rates depending on  the  form of algae present  (Feminella and Resh  1991).    In  the present  study,  there were  clear opposing  influences of  the  two main  growth  forms measured, with general patterns suggesting diatom cover generated a  positive  response,  and  filamentous  cover  a  negative  response,  of  invertebrate  communities.   However,  invertebrate  community  composition  did  not  vary  strongly  between high and  low productivity streams.   Differences were mainly due to changes  in the densities of the three dominant taxa: the blackfly  larvae Prosimulium spp., the  mayfly Baetis  spp.  and  amphipod Echinogammarus  spp..   Barquin  and Death  (2004)  also found these taxa in high densities in streams of this region.      Percent diatom cover was the best predictor of diversity with both taxonomic  and  rarefied  richness  increasing  log‐linearly  as  cover  increased.    This  mirrors  the  hypothesis  we  set  of  a  log‐linear  curve  for  the  relationship  between  primary  productivity  and  diversity  that  several  recent  studies  have  found  in  benthic  communities  (Death 2002, Barquin 2004, Tonkin and Death  In prep., Tonkin et al.  In  prep.).    Diatom  forms  of  algae  are  the  most  important  food  source  for  a  high  Chapter 4: Productivity‐diversity and periphyton assemblages      110    proportion of benthic  invertebrates as grazers  tend  to be able  to assimilate diatoms  better than other algal growth forms (Lamberti et al. 1989).    The  percent  of  EPT  animals  responded  negatively  at  the  highest  levels  of  diatom cover.   These  taxa are considered sensitive  to “pollution”,  thus  the  response  found  here  may  indicate  that  higher  levels  of  diatom  cover  are  associated  with  nutrient enrichment.  However, the response of % EPT at high diatom cover is largely  due to one dominant taxon – Baetis spp. – which was one of the most abundant taxa in  all  streams.    The decline  at higher diatom  levels may be due  to  the  fact  that more  palatable  forms  of  diatoms  become  shaded  out  by  prostrate  forms  when  growth  becomes more prolific.  Baetis often use diatoms as a food source and their prevalence  can vary greatly with types of algae depending on the stage in their lifecycle (Dudley et  al. 1986).  Mayflies generally tend to favour grazing diatoms (Jacoby 1987) and due to  the large proportion of grazers, EPT taxa often respond negatively to filamentous algae  (Quinn and Hickey 1990, Suren 2005).  Consequently, we expected a strong decline in  % EPT with  filamentous cover but  it remained relatively high up to three quarters of  bed cover of filamentous algae.  Filamentous algae may provide a greater retention of  organic detrital material that in turn would support more shredding taxa.  However the  shredder/predator  Echinogammarus  spp.  did  not  respond  to  either  form  of  periphyton.    Barquin  and  Death  (2004)  found  Echinogammarus  to  be  dominant  in  spring‐fed streams in Cantabria and suggested this may be a consequence of increased  biotic interactions associated with environmental stability (Connell 1978), thus causing  suppression of other  invertebrates.   We found no evidence to suggest  lower diversity  at sites with greater densities of Echinogammarus.    The number of  animals  increased  rapidly with  filamentous  cover  largely  as  a  result  of  Prosimulium  spp.  becoming  dominant.    The  response  of  Prosimulium  to  increased filamentous cover is variable due to their life histories (Towns 1981, Dudley  et al. 1986, Morin and Peters 1988); black flies (Simuliidae) are often associated with  bare  substrates but  smaller  individuals are often  found  in high densities attached  to  filamentous algae (Dudley et al. 1986).  Black fly larvae are filter feeders capturing their  food from drifting organic seston and so do not directly browse upon periphyton, but  can  be  found  in  high  densities  in  enriched  rivers  feeding  upon  drifting  algal  cells  Chapter 4: Productivity‐diversity and periphyton assemblages      111    (Peterson  et  al.  1985)  as  do  other  filter‐feeding  invertebrates  (Benke  and Wallace  1980, Wallace and Merritt 1980).      The  number  of  taxa  inhabiting  substrates  did  not  change  with  increasing  filamentous  algae  cover.    However,  due  to marked  increases  in  Prosimulium  spp.,  rarefied richness declined strongly at  intermediate  levels.   When  levels of periphyton  reach  greater  densities  and  epilithic  diatoms  are  replaced  by  macroalgae  such  as  filamentous green algae, more complex relationships tend to develop involving habitat  provision and exclusion as well as direct food interactions (Dudley et al. 1986).  Dense  layers of filamentous algae can lead to the displacement of sensitive taxa by those that  can tolerate  large diurnal fluctuations  in dissolved oxygen.   For example, chironomids  are often associated with macroalgae whereas EPT taxa are not (Power 1990, Koksvik  and Reinertsen 2008, Maasri et al. 2008).   Although we have coarsely classed growth  forms  into  two broad  categories, benthic  invertebrates often have  specialist  feeding  traits for these two algal forms.  Steinman et al. (1992) found the snail Elimia avoided  both  gelatinous  and  unbranched  filamentous  growth  forms  and  favoured  diatoms;  they suggest this may be due to a  lack of ability to digest    filamentous  forms.    It has  also been  suggested  that  stream herbivores  typically  reject  filamentous  algae when  selecting food sources (Gregory 1983), possibly because they often have high cellulose  content and thick walls (Lamberti and Moore 1984).   This selectivity goes beyond the  coarse  levels  assessed  here.    For  example, Hambrook  and  Sheath  (1987)  show  that  selective grazing can come within  filamentous  forms where preferences were shown  for simple unbranched rhodophytes over branched and mucilaginous.      The dense levels of filamentous algae found here is potentially a result of such  selective grazing which can alter plant communities by reducing palatable  forms and  leaving  non‐palatable  forms  behind  (Sousa  1984, Wardle  et  al.  2001, Brathen  et  al.  2007).  The dominant invertebrate taxa found in this study are not typically associated  with grazing  filamentous algae, although as previously stated, simuliids can be  found  filter  feeding on drifting  algal  cells but  this does not  suppress  growth.   Grazers  can  alter  periphyton  community  structure  within  and  between  micro  and  macro  algal  forms  in  lotic  systems  (Lamberti and Resh 1983, McAuliffe 1984, Dudley et al. 1986,  Feminella  and  Resh  1991),  and  the  suppression  of  palatable  forms  can  lead  to  Chapter 4: Productivity‐diversity and periphyton assemblages      112    communities consisting of resistant prostrate blue‐green algae (Hart 1985, Power et al.  1988).   In fact, selectivity  is often so strong that grazing  insects may remove, without  ingesting, unfavourable forms in order to allow favourable forms to remain for grazing  (Hart 1985).      Management implications   Many relationships in nature are nonlinear (Dodds et al. 2010), thus we must consider  this when  interpreting  results  such  as  those  found  here.   Nonlinearity  has  recently  been  recognised  for  the  setting  of  ecological  thresholds  of  anthropogenic  impact  (Hilderbrand et al. 2010) and assessing recovery from stressors (Clements et al. 2010)  in lotic systems.  Odum et al. (1979) introduced the concept of a subsidy‐stress effect  to demonstrate  the  response of biotic  communities  to  environmental perturbations  such  as  increased  nutrients.    In  some  cases  an  ecosystem  may  initially  respond  positively to a perturbation, such as  increased nutrients, before becoming stressed at  greater levels.  Although shifts in the form of periphyton cover is not an environmental  perturbation  per  se,  if we  consider  this within  a  subsidy‐stress  framework  (i.e.  the  environmental gradient as  the percent  cover of periphyton  forms and  the biological  response  as  invertebrate  community  metrics),  increased  diatom  cover  generally  appears to be a subsidy and increased filamentous cover a stressor.  This subsidy‐stress  effect is common in natural systems (Odum et al. 1979, Megonigal et al. 1997, King and  Richardson 2007) and can occur in lotic systems as a result of land‐use change altering  levels of nutrients, available  light and sediment (Quinn 2000, Niyogi et al. 2007).   It  is  likely that the real perturbation leading to these subsidy‐stress effects in these streams  is  increasing  nutrient  levels.    Long  term  increases  in  nutrients  can  lead  to  the  proliferations  of  filamentous  algae  similar  to  those  found  here.    Therefore,  growth  forms of periphyton are most likely reflecting long‐term nutrient levels and leading to  individual subsidy and stress effects on diversity.    The need  for  thresholds  in  the management of  lotic  systems  (Clements et al.  2010, Dodds et al. 2010, Hilderbrand et al. 2010), leads us to suggest filamentous algal  cover would be a useful  threshold  indicator.   Approximately 40 %  filamentous cover  led to a decline in diversity (richness and rarefied richness) whereas EPT taxa remained  Chapter 4: Productivity‐diversity and periphyton assemblages      113    relatively constant up to approximately 75 % cover.   Welch et al. (1988) found  levels  greater than 20 % filamentous algae cover were considered nuisance levels.      The fact that diversity responded more strongly to the form of periphyton than  chlorophyll  a  implies  this  could  be  a more  useful  tool  for managers  interested  in  preserving  biodiversity.    Typically,  biomonitoring  involves  assessing  periphyton  biomass using either chlorophyll a and/or ash‐free dry weight (AFDW).  Although they  are  often  highly  correlated,  they  do  not  always  respond  in  the  same  way  to  environmental conditions (Biggs and Hickey 1994, Feminella and Hawkins 1995, Jowett  and Biggs 1997), and  they  can have very different  chlorophyll a:carbon  ratios  (Biggs  2000).    Rather  than  simply  measuring  chlorophyll  a  or  AFDW  of  periphyton  communities,  it may  be more  appropriate  to  assess  the  relative  abundance  of  the  different growth forms of periphyton.  Many have used components of the periphyton  community as an index for biomonitoring environmental conditions (Kelly and Whitton  1995, Pan et al. 1996, Kelly 1998b, a, Kelly and Whitton 1998, Hill et al. 2000), however  most  of  these  use  diatoms  exclusively  and  require  large  investments  in  time  and  money to complete.  We have found that periphyton community composition does not  need to be measured to a  low taxonomic  level, and rapid assessment protocols have  been  developed  for  this  assessment  (Biggs  and  Kilroy  2000).   Often  the  distinction  between  communities  with  shifts  in  relative  abundance  of  filamentous  algae  and  diatoms is obvious.  Suren et al. (2003) found invertebrate communities changed from  insect to non‐insect dominated with shifts  in the periphyton community from diatom  to filamentous algae dominated.     Conclusions  Our  research  found  no  relationship  between  a  commonly  used  surrogate  of  productivity  (chlorophyll a) and diversity  in these streams.   This  is a common  finding  when exploring productivity‐diversity relationships at smaller spatial scales (Waide et  al. 1999, Mittelbach et al. 2001).   As hypothesized, this relationship was underpinned  by contrasting  responses of  invertebrate communities  to  the  form of  the periphyton  community.   General patterns suggest that diversity responds positively to  increasing  diatom cover, however sensitive EPT taxa decline at  intermediate  levels.   Conversely,  Chapter 4: Productivity‐diversity and periphyton assemblages      114    invertebrate communities responded negatively to increasing filamentous algae cover.   This  underlying  response  to  the  form  of  periphyton  community  provides  some  explanation  for  the  lack of a clear  relationship between productivity and diversity  in  lotic  systems.    Additionally,  our  findings  have  important management  implications  which demonstrate that the rapid assessment of periphyton growth form can be highly  relevant  to higher  trophic  levels.   We suggest a  threshold of 40 %  filamentous cover  could be set for managers wishing to minimise deleterious effects of eutrophication on  mmunities.   associated invertebrate co   References  Abrams,  P. A.  1995. Monotonic  or  unimodal  diversity  productivity  gradients  ‐ What  does competition theory predict. Ecology 76:2019‐2027.  Allan, J. D. 1995. Stream Ecology: Structure and Function of Running Waters. Chapman  and Hall, London.  Barquin, J. 2004. Spatial patterns of invertebrate communities in spring and runoff‐fed  streams.  PhD  thesis.  PhD  thesis, Massey  University,  Palmerston  North,  New  Zealand.  Barquin,  J., and R. G. Death. 2004. Patterns of  invertebrate diversity  in  streams and  freshwater springs in Northern Spain. Archiv Fur Hydrobiologie 161:329‐349.  Benke, A. C., and J. B. Wallace. 1980. Trophic basis of production among net‐spinning  caddisflies in a souther Appalachian stream. Ecology 61:108‐118.  Biggs,  B.  J.  F.  2000.  New  Zealand  Periphyton  Guideline:  Detecting, Monitoring  and  Managing of Enrichment of Streams. Ministry for the Environment, Wellington,  New Zealand.  Biggs, B. J. F., and C. W. Hickey. 1994. Periphyton responses to a hyrdaulic‐gradient in a  regulated river in New Zealand. Freshwater Biology 32:49‐&.  Biggs,  B.  J.  F.,  and  C.  Kilroy.  2000.  Stream  periphyton  monitoring  manual.  NIWA,  Christchurch.  Chapter 4: Productivity‐diversity and periphyton assemblages      115    Brathen, K. A., R. A. Ims, N. G. Yoccoz, P. Fauchald, T. Tveraa, and V. H. Hausner. 2007.  Induced  shift  in ecosystem productivity ? Extensive  scale effects of abundant  large herbivores. Ecosystems 10:773‐789.  Chase,  J.  M.,  and  M.  A.  Leibold.  2002.  Spatial  scale  dictates  the  productivity‐ biodiversity relationship. Nature 416:427‐430.  Clarke,  K.  R.  1993.  Nonparametric  multivariate  analyses  of  changes  in  community  structure. Australian Journal of Ecology 18:117‐143.  Clarke,  K.  R.,  and  R. N. Gorley.  2006.  PRIMER  v6: User Manual/Tutorial.  PRIMER‐E,  Plymouth.  Clements, W.  H.,  N.  K. M.  Vieira,  and  D.  L.  Sonderegger.  2010.  Use  of  ecological  thresholds  to  assess  recovery  in  lotic  ecosystems.  Journal  of  the  North  American Benthological Society 29:1017‐1023.  Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs. Science 199:1302‐ 1310.  Currie, D. J. 1991. Energy and large‐scale patterns of animal‐species and plant‐species  richness. American Naturalist 137:27‐49.  Currie,  D.  J.,  and  V.  Paquin.  1987.  Large‐scale  biogeographical  patterns  of  species  richness of trees. Nature 329:326‐327.  Death, R. G. 2002. Predicting invertebrate diversity from disturbance regimes in forest  streams. Oikos 97:18‐30.  Death, R. G. 2008. Effects of floods on aquatic invertebrate communities.in J. Lancaster  and R. A. Briers, editors. Insects: Challenges to Populations. CAB International,  UK, pp 103‐121.  Death,  R.  G.,  and  E. M.  Zimmermann.  2005.  Interaction  between  disturbance  and  primary  productivity  in  determining  stream  invertebrate  diversity.  Oikos  111:392‐402.  Dodds, W.  K., W.  H.  Clements,  K.  Gido,  R.  H.  Hilderbrand,  and  R.  S.  King.  2010.  Thresholds,  breakpoints,  and  nonlinearity  in  freshwaters  as  related  to  management. Journal of the North American Benthological Society 29:988‐997.  Chapter 4: Productivity‐diversity and periphyton assemblages      116    Dudley, T. L., S. D. Cooper, and N. Hemphill. 1986. Effects of Macroalgae on a Stream  Invertebrate Community. Journal of the North American Benthological Society  5:93‐106.  Feminella, J. W., and C. P. Hawkins. 1995. Interactions between stream herbivores and  periphyton: A quantitative analysis of past experiments.  Journal of  the North  American Benthological Society 14:465‐509.  Feminella,  J.  W.,  and  V.  H.  Resh.  1991.  Herbivorous  caddisflies,  macroalgae,  and  epilithic  microalgae  ‐  dynamic  interactions  in  a  stream  grazing  system.  Oecologia 87:247‐256.  Fisher, S. G., L. J. Gray, N. B. Grimm, and D. E. Busch. 1982. Temporal Succession  in a  Desert  Stream  Ecosystem  Following  Flash  Flooding.  Ecological  Monographs  52:93‐110.  Fukami, T., and P. J. Morin. 2003. Productivity‐biodiversity relationships depend on the  history of community assembly. Nature 424:423‐426.  Gaston, K. J. 2000. Global patterns in biodiversity. Nature 405:220‐227.  Graham, A. A., D. J. McCaughan, and F. S. McKee. 1988. Measurement of surface area  of stones. Hydrobiologia 157:85‐87.  Gregory, S. V. 1983. Plant‐herbivore interactions in stream systems. Pages 157‐189 in J.  R. Barnes and G. W. Minshall, editors. Stream ecology: application and testing  of general ecological theory, Plenum Press, New York.  Gresens,  S.  E.,  and  R.  L.  Lowe.  1994.  Periphyton  Patch  Preference  in  Grazing  Chironomid Larvae. Journal of the North American Benthological Society 13:89‐ 99.  Grime,  J. P. 1973. Competitive exclusion  in herbaceous  vegetation. Nature 242:344‐ 347.  Hall, M., E. Frank, G. Holmes, B. Pfahringer, P. Reutmann, and I. H. Witten. 2009. The  WEKA  Data Mining  Software:  An  Update.  SIGKDD  Explorations,  Volume  11,  Issue 1.  Hambrook, J. A., and R. G. Sheath. 1987. Grazing of fresh‐water Rhodophyta. Journal of  Phycology 23:656‐662.  Chapter 4: Productivity‐diversity and periphyton assemblages      117    Hart,  D.  D.  1985.  Grazing  insects  mediate  algal  interactions  in  a  stream  benthic  community. Oikos 44:40‐46.  Hilderbrand, R. H., R. M. Utz, S. A. Stranko, and R. L. Raesly. 2010. Applying thresholds  to forecast potential biodiversity loss from human development. Journal of the  North American Benthological Society 29:1009‐1016.  Hill, B. H., A. T. Herlihy, P. R. Kaufmann, R.  J. Stevenson, F. H. McCormick, and C. B.  Johnson.  2000.  Use  of  periphyton  assemblage  data  as  an  index  of  biotic  integrity. Journal of the North American Benthological Society 19:50‐67.  Hillebrand,  H.  2009. Meta‐analysis  of  grazer  control  of  periphyton  biomass  across  aquatic ecosystems. Journal of Phycology 45:798‐806.  Hillebrand,  H.,  and  B.  J.  Cardinale.  2010.  A  critique  for  meta‐analyses  and  the  productivity‐diversity relationship. Ecology 91:2545‐2549.  Hurlbert, S. H. 1971. The non‐concept of  species diversity: a critique and alternative  parameters. Ecology 52:577‐586.  Huston, M. 1979. A general hypothesis of  species diversity. The American Naturalist  113:81‐100.  Huston,  M.  1994.  Biological  Diversity:  The  Coexistence  of  Species  on  Changing  Landscapes. Cambridge University Press, Cambridge, UK.  Jacoby, J. M. 1987. Alterations in periphtyon characteristics due to grazing in a cascade  foothill stream. Freshwater Biology 18:495‐508.  Jowett, I. G., and B. J. F. Biggs. 1997. Flood and velocity effects on periphyton and silt  accumulation  in  two New Zealand  rivers. New Zealand  Journal of Marine and  Freshwater Research 31:287‐300.  Jowett,  I.  G.,  and  J.  Richardson.  1990.  Microhabitat  preferences  of  benthic  invertebrates  in a New Zealand river and  the development of  in‐stream  flow‐ habitat  models  for  Deleatidium  spp.  New  Zealand  Journal  of  Marine  and  Freshwater Research 24:19‐30.  Kelly, M.  G.  1998a.  Use  of  community‐based  indices  to monitor  eutrophication  in  European rivers. Environmental Conservation 25:22‐29.  Kelly, M. G. 1998b. Use of the trophic diatom index to monitor eutrophication in rivers.  Water Research 32:236‐242.  Chapter 4: Productivity‐diversity and periphyton assemblages      118    Kelly,  M.  G.,  and  B.  A.  Whitton.  1995.  Trophic  diatom  index  ‐  a  new  index  for  monitoring eutrophication in rivers. Journal of Applied Phycology 7:433‐444.  Kelly, M. G., and B. A. Whitton. 1998. Biological monitoring of eutrophication in rivers.  Hydrobiologia 384:55‐67.  King, R. S., and C.  J. Richardson. 2007. Subsidy‐stress  response of macroinvertebrate  community  biomass  to  a  phosphorus  gradient  in  an  oligotrophic  wetland  ecosystem. Journal of the North American Benthological Society 26:491‐508.  Koksvik, J. I., and H. Reinertsen. 2008. Changes in macroalgae and bottom fauna in the  winter period  in  the  regulated Alta River  in northern Norway. River Research  and Applications 24:720‐731.  Kondoh, M.  2001. Unifying  the  relationships  of  species  richness  to  productivity  and  disturbance.  Proceedings  of  the  Royal  Society  of  London  Series  B‐Biological  Sciences 268:269‐271.  Lake, P. S. 2000. Disturbance, patchiness, and diversity in streams. Journal of the North  American Benthological Society 19:573‐592.  Lamberti, G. A., S. V. Gregory, L. R. Ashkenas, A. D. Steinman, and C. D. McIntire. 1989.  Productive  capacity  of  periphyton  as  a  determinant  of  plant  herbivore  interactions in streams. Ecology 70:1840‐1856.  Lamberti, G. A., and J. W. Moore. 1984. Aquatic  insects as primary consumers. Pages  164‐195  in V. H. Resh  and D. M.  Rosenberg,  editors.  The  ecology  of  aquatic  insects, Praeger, New York.  Lamberti, G. A.,  and V. H. Resh. 1983.  Stream periphyton  and  insect herbivores:  an  experimental study of grazing by a caddisfly population. Ecology 64:1124‐1135.  Leibold,  M.  A.  1999.  Biodiversity  and  nutrient  enrichment  in  pond  plankton  communities. Evolutionary Ecology Research 1:73‐95.  Lenat, D. R. 1988. Water quality assessment of streams using a qualitative collection  method for benthic invertebrates. Journal of the North American Benthological  Society 7:222‐233.  Maasri,  A.,  S.  Fayolle,  E.  Gandouin,  R.  Garnier,  and  E.  Franquet.  2008.  Epilithic  chironomid  larvae  and water  enrichment:  is  larval  distribution  explained  by  Chapter 4: Productivity‐diversity and periphyton assemblages      119    epilithon  quantity  or  quality?  Journal  of  the  North  American  Benthological  Society 27:38‐51.  McAuliffe,  J.  R.  1984.  Resource  depression  by  a  stream  herbivore  ‐  effects  on  distributions and abundances of other grazers. Oikos 42:327‐333.  Megonigal,  J.  P., W.  H.  Conner,  S.  Kroeger,  and  R.  R.  Sharitz.  1997.  Aboveground  production  in  Southeastern  floodplain  forests:  A  test  of  the  subsidy‐stress  hypothesis. Ecology 78:370‐384.  Mittelbach, G. G., C. F. Steiner, S. M. Scheiner, K. L. Gross, H. L. Reynolds, R. B. Waide,  M. R. Willig, S. I. Dodson, and L. Gough. 2001. What is the observed relationship  between species richness and productivity? Ecology 82:2381‐2396.  Morin, A., W. Lamourex, and  J. Busnarda. 1999. Empirical models predicting primary  productivity  from chlorophyll a and water temperature  for stream periphyton  and  lake  and  ocean  phytoplankton.  Journal  of  the  North  American  Benthological Society 18:299‐307.  Morin, A., and R. H. Peters. 1988. Effect of microhabitat features, seston quality, and  periphyton on abundance of overwintering black fly larvae in southern Quebec.  Limnology and Oceanography 33:431‐446.  Niyogi, D. K., M. Koren, C. J. Arbuckle, and C. R. Townsend. 2007. Stream communities  along  a  catchment  land‐use  gradient:  subsidy‐stress  responses  to  pastoral  development. Environmental Management 39:213‐225.  Odum,  E. P.,  J.  T.  Finn,  and  E. H.  Franz.  1979.  Perturbation‐theory  and  the  subsidy‐ stress gradient. Bioscience 29:349‐352.  Pan, Y. D., R. J. Stevenson, B. H. Hill, A. T. Herlihy, and G. B. Collins. 1996. Using diatoms  as  indicators of ecological  conditions  in  lotic  systems: A  regional assessment.  Journal of the North American Benthological Society 15:481‐495.  Peterson, B.  J.,  J. E. Hobbie, A. E. Hershey, M. A.  Lock, T. E.  Ford,  J. R. Vestal, V.  L.  McKinley, M. A.  J. Hullar, M. C. Miller, R. M. Ventullo,  and G.  S. Volk.  1985.  Transformation of a Tundra River from Heterotrophy to Autotrophy by Addition  of Phosphorus. Science 229:1383‐1386.  Pfankuch, D. J. 1975. Stream Reach  Inventory and Channel Stability Evaluation. USDA  Forest Service, Region 1, Missoula, Montana.  Chapter 4: Productivity‐diversity and periphyton assemblages      120    Power, M. E. 1990. Benthic Turfs Vs Floating Mats of Algae in River Food Webs. Oikos  58:67‐79.  Power, M. E., A.  J. Stewart, and W. J. Matthews. 1988. Grazer Control of Algae  in an  Ozark Mountain  Stream  ‐  Effects  of  Short‐Term  Exclusion.  Ecology  69:1894‐ 1898.  Quinn, J. M. 2000. Effects of pastoral development. Pages 208‐229  in K. J. Collier and  M.  J. Winterbourn,  editors.  New  Zealand  stream  invertebrates:  ecology  and  implications for management. The Caxton Press, Christchurch, New Zealand.  Quinn,  J. M.,  and  C. W. Hickey.  1990.  Characterisation  and  classification  of  benthic  invertebrate  communities  in  88  New  Zealand  rivers  in  relation  to  environmental  factors.  New  Zealand  Journal  of  Marine  and  Freshwater  Research 24:387‐409.  Resh, V. H., A. V. Brown, A. P. Covich, M. E. Gurtz, H. W. Li, G. W. Minshall, S. R. Reice,  A. L. Sheldon, J. B. Wallace, and R. C. Wissmar. 1988. The role of disturbance in  stream  ecology.  Journal  of  the North  American  Benthological  Society  7:433‐ 455.  Rosenzweig, M. L. 1992. Species‐diversity gradients ‐ we know more and less than we  thought. Journal of Mammalogy 73:715‐730.  Rosenzweig, M.  L.  1995.  Species Diversity  in  Space  and  Time.  Cambridge University  Press, Cambridge, UK.  Sanders,  H.  L.  1968.  Marine  benthic  diversity:  a  comparative  study.  American  Naturalist 102:243‐282.  Sousa, W. P. 1984. Intertidal mosaics ‐ patch size, propagule availability, and spatially‐ variable patterns of succession. Ecology 65:1918‐1935.  Steinman, A. D.,  and G. A.  Lamberti.  1996. Biomass  and  pigments  of  benthic  algae.  Pages 295‐314  in F. R. Hauer and G. A. Lamberti, editors. Methods  in Stream  Ecology. Academic Press, San Diego, CA.  Steinman, A. D., P. J. Mulholland, and W. R. Hill. 1992. Functional‐responses associated  with growth form in stream algae. Journal of the North American Benthological  Society 11:229‐243.  Chapter 4: Productivity‐diversity and periphyton assemblages      121    Suren, A. M. 2005. Effects of deposited  sediment on patch  selection by  two grazing  stream invertebrates. Hydrobiologia 549:205‐218.  Suren, A. M., B.  J.  F. Biggs, M.  J. Duncan,  L. Bergey,  and P.  Lambert.  2003. Benthic  community  dynamics  during  summer  low‐flows  in  two  rivers  of  contrasting  enrichment 2.  Invertebrates. New  Zealand  Journal of Marine  and  Freshwater  Research 37:71‐83.  Suren, A. M.,  and  T. Riis.  2010.  The  effects of plant  growth on  stream  invertebrate  communities  during  low  flow:  a  conceptual  model.  Journal  of  the  North  American Benthological Society 29:711‐724.  Tachet, H., P. Richoux, M. Bournaud, and P. Usseglio‐Polatera. 2000. Invertebres d'eau  douce. Systematique, biologie, ecologie. CNRS Editions, Paris.  Tonkin,  J.  D.,  and  R.  G.  Death.  In  prep.  Consistent  effects  of  productivity  and  disturbance on diversity between landscapes.  Tonkin,  J. D., R. G. Death, and K.  J. Collier.  In prep. Do productivity and disturbance  interact to modulate macroinvertebrate diversity in streams?  Towns, D. R. 1981. Effects of  artificial  shading on periphyton  and  invertebrates  in  a  New Zealand stream. New Zealand Journal of Marine and Freshwater Research  15:185‐192.  Waide, R. B., M. R. Willig, C.  F.  Steiner, G. Mittelbach,  L. Gough,  S.  I. Dodson, G. P.  Juday,  and  R.  Parmenter.  1999.  The  relationship  between  productivity  and  species richness. Annual Review of Ecology and Systematics 30:257‐300.  Wallace, J. B., and R. W. Merritt. 1980. Filter‐feeding ecology of aquatic insects. Annual  Review of Entomology 25:103‐132.  Wardle, D. A., G. M. Barker, G. W. Yeates, K. I. Bonner, and A. Ghani. 2001. Introduced  browsing  mammals  in  New  Zealand  natural  forests:  Aboveground  and  belowground consequences. Ecological Monographs 71:587‐614.  Welch,  E. B.,  J. M.  Jacoby, R. R. Horner,  and M. R.  Seeley.  1988. Nuisance  biomass  levels of periphytic algae in streams. Hydrobiologia 157:161‐168.  Whittaker, R. J. 2010. Meta‐analyses and mega‐mistakes: calling time on meta‐analysis  of the species richness‐productivity relationship. Ecology 91:2522‐2533.  Chapter 4: Productivity‐diversity and periphyton assemblages      122    Whitton, B. A., and M. G. Kelly. 1995. Use of algae and other plants  for monitoring  rivers. Australian Journal of Ecology 20:45‐56.  Wolman, M. J. 1954. A method of sampling coarse river bed material. Transactions of  the American Geophysical Union 35:951‐956.          Chapter 4: Productivity‐diversity and periphyton assemblages      123    Appendix  1: Mean physicochemical  variables  and  site  characteristics  recorded  at  24  streams  in  the Cantabrian Mountains,  Spain,  in  July  2007.   Alt.  =  altitude, Pfank. Bottom = the bottom component of the Pfankuch stability index, Cond. = conductivity, Temp. = temperature.                                              Site  Name  Catchment  Longitude  Latitude  Alt.  Width  (m)  Depth  (cm)  Velocity  (m/s)  Substrate  size index  Pfank.  bottom  pH  Cond.  (µs/cm)  Temp.  (°C)  1  Rio Besaya @ Helguera  Besaya  ‐4.03276  43.158628  214  10.34  27  0.49  123.07  23  9.3  371  15.49  2  Rio Eracia @ Helguera  Besaya                                              ‐4.030378  43.159622  222  7.60  23  0.36  141.54  29  9.03  166  15.91  3  Rio Bisuena @ Barcena Pie de Concha  Besaya ‐4.06826  43.121292   309  4.31  26  0.56  131.88  39  8.7  269  14.7  4  Torina @ Barcena Pie de Concha  Besaya ‐4.054722  43.126793   299  7.30  22.8  0.20  142.69  27  9.19  124  14.96  5  Santiurde @ Santiurde  Besaya ‐4.078611  43.063231   634  4.45  15.2  0.48  127.15  44  8.99  221  12.73  6  Rumadero @ Pesquera  Besaya ‐4.074784  43.083177   574  2.10  12.6  0.19  132.02  41  8.84  329  13.77  7  Rio Leon @ San Martin de Quevedo  Besaya ‐4.038779  43.139996   264  5.13  18  0.35  142.14  43  8.76  193  16  8  Ayo Valdeiguna @ Pedredo  Besaya ‐4.067773  43.192911   185  6.13  16.4  0.23  152.16  26  8.65  205  16.55  9  Rio Argoza @ Barcena Mayor  Saja ‐4.232753  43.156963   422  17.50  30.2  0.63  178.53  27  9.02  127  15.2  10  Rio Saja @ Renedo  Saja ‐4.30455  43.194076   293  30.70  23.4  0.56  145.22  42  8.97  130  17.79  11  Rio de la Magdalena @ San Andres  Pas ‐3.897027  43.096378  412  9.63  23.2  0.31  166.81  30  9.04  194  12.77  12  Ayo Salcera @ San Miguel de Luena  Pas ‐3.899921  43.112736   347  2.67  13.4  0.46  158.08  19  8.85  216  12.94  13  Rio Viaña @ Viaña  Pas ‐3.804674  43.155633   326  6.97  19.4  0.47  139.45  24  8.66  81  18.63  14  Rio Pas @ Pandillo  Pas ‐3.759107  43.164859  366  9.50  20.6  0.54  192.32  30  8.77  130  18.36  15  Llerana @ Coterillo  Pisuena ‐3.794589  43.263832  207  5.37  18.4  0.38  137.48  32  8.44  256  15.38  16  Rio Pisuena @ Barcena de Carriedo  Pisuena ‐3.823485  43.241998   181  7.33  17.2  0.39  102.32  45  8.48  283  16.82  17  Nansa @ Puentenansa  Nansa ‐4.406832  43.257111  168  9.47  20.8  0.36  121.21  41  8.75  303  15.41  18  Ayo de Hoyamala @ Puentenansa  Nansa ‐4.406179  43.257101   163  6.15  16.8  0.33  110.22  33  8.59  280  14.02  19  Lamasón @ Quintanilla  Nansa ‐4.476253  43.25448   238  6.40  17  0.35  112.50  44  8.78  144  16.42  20  Ayo de Traveseras @ Quitanilla  Nansa ‐4.30588  43.256427   239  4.00  13.6  0.22  196.25  24  8.76  247  15.13  21  Palomba @ Paracuelles  Ebro ‐4.211634  43.019856   895  1.93  14.4  0.24  61.12  20  9.24  402  12.99  22  Rio Hijar @ Espanilla  Ebro ‐4.226527  43.019571   937  13.00  13.4  0.45  119.91  43  9.15  121  17.17  23  Rio Rucebos @ Soto  Ebro ‐4.222249  43.035916   960  5.00  14.6  0.52  131.04  34  9.06  68  13.01  24  Rio Guares @ Abiado  Ebro ‐4.289066  43.016011   1061  2.27  15.2  0.47  173.46  26  8.63  83  11.55  Chapter 5  Global patterns in stream diversity – productivity relationships                  Chapter 5: Global productivity‐diversity relationships      Abstract  125    More productive environments typically have more species, although the specific form  of  this  relationship  is  unclear  and  varies  with  spatial  scale.    This  relationship  has  received little direct attention in lotic systems and thus there is no clear understanding  of the form of the trend.  To examine this relationship in lotic systems and whether a  universal trend exists, we examined and compared 24 streams of Cantabria, Spain and  24 of the central North  Island, New Zealand between February and July 2007.   Based  on the notion that productivity dictates the upper  limit to richness  in streams rather  than  to  increase  the  effect  of  competitive  interactions,  we  predicted  a  log‐linear  relationship  in  both  regions.    Productivity,  assessed  as  chlorophyll  a,  was  approximately  three  times  higher  in  Spanish  streams  but  taxonomic  richness  of  invertebrates  did  not  differ  between  the  two  regions.    Taxonomic  richness  (and  Shannon  diversity)  only  responded  to  productivity  in  the  New  Zealand  streams,  exhibiting  the  predicted  log‐linear  increase.    In  fact,  the  only metric  to  respond  to  productivity  in  the  Spanish  streams  was  the  number  of  animals  which  increased  monotonically.    However,  this  relationship  was  assessed  at  different  scales  of  productivity  between  the  two  regions.   When  fit  to  the  same  axis,  richness  in  the  Spanish  streams  was  simply  a  continuation  of  the  log‐linear  curve  fit  to  the  New  Zealand streams.   With the  large range of productivity assessed and no evidence of a  decline  in diversity with higher productivity,  it appears  that  the view of productivity  setting maximum  richness,  rather  than  leading  to  stronger  competitive  interactions,  may be global phenomenon in streams.           Chapter 5: Global productivity‐diversity relationships      Introduction  126    Predicting  diversity  remains  a  fundamental  challenge  in  community  ecology  (e.g.  Huston  1994,  Hubbell  2001,  Ricklefs  2004).    Global  patterns  suggest  diversity  is  a  function  of  a  few  broad‐scale  factors  such  as  latitude,  precipitation,  temperature,  altitude and  land mass (Gaston 2000).   Of the multitude of factors that can  influence  diversity  (Palmer  1994,  Vinson  and  Hawkins  1998),  disturbance  (Connell  1978,  Wootton  1998,  Lake  2000)  and  productivity  (Currie  1991,  Waide  et  al.  1999,  Mittelbach et al. 2001) prevail, although universal patterns remain elusive (Waide et al.  1999, Mackey and Currie 2001, Mittelbach et al. 2001).   What  is  likely  is disturbance  and productivity  interact to affect diversity but the nature of this  interaction remains  unclear (Sousa 1984, Pickett and White 1985, Currie 1991, Wootton 1998, Waide et al.  1999, Mittelbach  et  al.  2001, Death  and  Zimmermann  2005,  Cardinale  et  al.  2006).    Few systems are more heavily influenced by disturbance than streams (Resh et  al. 1988, Lake 2000, Death 2008), but the response of diversity to disturbance is likely  to  be  dependent  on  productivity  (Death  2002,  Cardinale  et  al.  2005,  Death  and  Zimmermann 2005, Cardinale et  al. 2006).    Several empirical  studies have  found  an  interaction  between  productivity  and  disturbance  to  be  important  in  controlling  diversity  (e.g.  Wilson  and  Tilman  2002,  Kneitel  and  Chase  2004,  Death  and  Zimmermann 2005, Scholes et al. 2005).  However, due to the highly mobile nature of  stream  invertebrates,  successful  applications  of  common  models  such  as  the  Intermediate Disturbance Hypothesis  (IDH)  (Grime 1973b, Connell 1978, Sousa 1979)  and the Dynamic Equilibrium Model (DEM) (Huston 1979, 1994, Kondoh 2001) are rare  (but see Townsend et al. 1997).   What makes streams especially unique  is the strong  association between disturbance and productivity (Lake 2000, Death 2002, Tonkin and  Death  In  prep.‐a,  Tonkin  et  al.  In  prep.‐b).    Furthermore,  there  is  little  evidence  in  streams of a competition‐colonisation  trade‐off  (Reice 1985, Death and Winterbourn  1995).    Death  (2002)  proposed  a  model  suggesting  disturbance  simply  resets  the  colonization process  and productivity,  rather  than  competitive  interactions,  sets  the  upper  limit  to  richness  in  streams.    However,  whether  a  great  enough  range  of  productivity  was  sampled  to  elucidate  the  full  pattern  is  unclear  as  this  was  only  Chapter 5: Global productivity‐diversity relationships      127    assessed  in  New  Zealand  forest  streams.    In  order  to  understand  this  interaction  between  productivity  and  disturbance  in  depth,  clarification  of  the  productivity‐ diversity relationship in streams is required.    The  two most  common  forms  of  the  productivity‐diversity  relationship  in  all  systems  are  unimodal  (e.g.  Grime  1973a,  Huston  1979,  Rosenzweig  1992,  1995,  Leibold  1999, Waide  et  al.  1999, Mittelbach et  al.  2001)  and  linear  (e.g. Currie  and  Paquin 1987, Currie 1991, Abrams 1995, Waide et al. 1999, Gaston 2000, Mittelbach et  al.  2001)  increases  in  diversity  with  increasing  productivity.    One  of  the  major  influences on the  form of this relationship  is the spatial scale of  investigation  (Currie  1991, Chase and Leibold 2002, Tonkin and Death In Prep.‐b) and the type of ecosystem  and/or organisms considered (Waide et al. 1999, Mittelbach et al. 2001).  Smaller scale  studies commonly find unimodal relationships between productivity and diversity but  at greater  spatial  scales,  this pattern often exhibits monotonic  increases  in diversity  with productivity (Waide et al. 1999, Mittelbach et al. 2001, Chase and Leibold 2002).   There  are  many  mechanisms  postulated  to  lead  to  this  unimodal  relationship  (Rosenzweig and Abramsky 1993, Abrams 1995, Waide et al. 1999), but  these often  require some form of competitive trade‐off.  However, little effort has been placed on  elucidating this relationship  in  lotic systems.   The  few to  look specifically have  found  both  unimodal  (Death  and  Zimmermann  2005,  Cardinale  et  al.  2006)  and  log‐linear  (Death  2002,  Barquin  2004,  Tonkin  and  Death  In  prep.‐a,  Tonkin  et  al.  In  prep.‐b)  increases in diversity with increasing productivity.    Not only are the effects of productivity dependent on scale, but regional effects  such  as  the  potential  pool  of  colonizers  can  completely  alter  the  response  to  environmental  factors.    Two  studies  by  Barquin  and  Death  (2004,  2006)  found  contrasting responses of invertebrates to environmental stability between Spanish and  New Zealand streams.  Thus we set out to (i) examine whether the response of stream  invertebrate  diversity  to  productivity  differs  between  these  two  regions  at  similar  distances  from  the  equator  in  the  Northern  and  Southern  Hemispheres;  and  (ii)  whether  a  universal  relationship  exists  for  the  productivity‐diversity  relationship  by  combining both regions and re‐examining the nature of the link.  We hypothesise that  diversity will be higher  in the Spanish streams as they occur within a  large continent  Chapter 5: Global productivity‐diversity relationships      128    with a greater potential pool of colonists.   However, we expect  that the response of  diversity to productivity will be similar between the two regions despite the different  taxonomic  composition.    That  is,  we  hypothesise  that  diversity  will  increase  with  increasing productivity up to a point before tapering off.  We also assess whether the  percent  of  Ephemeroptera,  Plecoptera  and  Trichoptera  (%  EPT)  respond  differently  than typical diversity measures as these taxa are sensitive to ‘pollution’, and associated  algal  proliferations,  in  lotic  systems.    Finally,  due  to  the  strong  link  between  productivity and disturbance in streams, we explore the relationship between stability  and diversity in these streams and expect that this may explain any lack of pattern with  productivity.  By combining and comparing two regions at opposing sides of the planet  e can potentially provide some clarity to global patterns in the productivity‐diversity  ystems.    w relationship of lotic s   Methods  Study sites  Twenty four streams were sampled  in each of the Cantabrian Mountains of Northern  Spain and Tongariro National Park in the central North Island of New Zealand between  February and July 2007 (Appendix 1).  The Cantabrian Mountains span approx. 483 km  east  to  west  along  the  northern  coast  of  Spain  reaching  2,648 m  asl  at  Torre  de  Cerredo.   Average rainfall ranges from approx. 1,200 to 1,600 mm p.a. depending on  location within  the region.   Land‐use surrounding sampling sites varied  from Atlantic  deciduous forest consisting predominantly of oak (Quercus spp.) and European beech  (Fagus spp.) to pasture and small urban settlements.  Sites were selected from six river  catchments: Rio Besaya, Rio Saja, Rio Pas, Rio Pisueña, Rio Nansa, and Rio Ebro.  Except  for Rio Ebro, which drains into the Mediterranean Sea, all rivers flow into the Atlantic.   Altitude ranged from 163 to 1061 m asl and average channel width ranged from 1.9 to  30.7 m.      The Tongariro National Park  is dominated by the central volcanic massif of Mt  Ruapehu  (2,797 m asl), Mt Ngauruhoe  (2,287 m asl) and Mt Tongariro  (1,967 m asl)  Chapter 5: Global productivity‐diversity relationships      129    and the Tihia‐kakaramea volcanic massif to the north.  The Kaimanawa Ranges rise to  ~1,799 m asl to the east of the Tongariro National Park Vegetation within and around  the park varies from broadleaf‐podocarp, mixed beech‐podocarp, exotic Pinus radiata  plantation, native tussock and scrubland, to bare ground in the eastern rain‐shadow of  the three central volcanoes.  The north and west has an average rainfall of 1,800‐3,500  mm yr‐1 with the south and east only receiving around 1,100 mm yr‐1 due to the rain‐ shadow cast by the three mountains from the prevailing westerly winds.  Sites ranged  from 518 to 1158 m asl and average width of stream channels ranged from 1.4 to 30  m.      Sampling protocol  Macroinvertebrates  Macroinvertebrates were sampled by taking five 0.1 m2 Surber samples (250 µm mesh  in New  Zealand  samples  and 500 µm  in  Spanish  samples)  from  random  locations  in  riffles throughout ~50 m reaches at each site.   Samples were stored  in 10 % formalin  and  later  sieved  to  500  µm  and  identified  in  the  laboratory  to  the  lowest  possible  taxonomic  level.    New  Zealand  samples  were  identified  using  available  keys  (e.g.  Towns and Peters 1996, Winterbourn et al. 2000) and Spanish  samples were mostly  identified  to morphospecies,  however,  where  possible  these  were  identified  using  available keys (e.g. Tachet et al. 2000).  Four simple measures of diversity were used in  this study:   number of  individuals, number of taxa (richness), Shannon diversity  index  (H’),  and  the  percent  of  Ephemeroptera,  Plecoptera,  Trichoptera  (%  EPT  animals)  (Lenat 1988).      Periphyton  Productivity  was  not  directly  measured  but  was  estimated  from  measures  of  chlorophyll  a  from  stream  substrata.    Chlorophyll  a  and  primary  productivity  in  streams are highly correlated (Morin et al. 1999).  Tonkin and Death (In prep.‐a) found  a  strong  linear  link  between  chlorophyll  a  biomass  accumulation  on  tiles  and  chlorophyll a content of stones (r2 = 0.74).    Chapter 5: Global productivity‐diversity relationships      130    Five stones (α axis < 60 mm) were randomly collected from each riffle for later  extraction of chlorophyll a.  From here stones and tiles were kept cool and dark on ice  before being stored at ‐20˚C.  Photosynthetic pigments were extracted from stones by  submerging  each  stone  in  known  volumes  of  90  %  acetone  for  24  hours  at  5˚C.   Absorbances at 750, 665 and 664 nm were read on a Varian Cary 50 conc UV‐Visible  Spectrophotometer  (Varian  Australia  Pty  Ltd, Mulgrave,  Australia)  for  New  Zealand  samples  and  on  a  Beckman  Coulter  DU®  Series  700  UV/Vis  Scanning  Spectrophotometer  (Beckman  Coulter  Inc.,  Brea,  CA, USA)  for  Spanish  samples  and  converted  to  chlorophyll  a  and  phaeophytin  pigment  concentration  using  Steinman  and Lamberti (1996).  These were then corrected for stone surface area and halved to  account for just the active growing area of the stone using Graham et al. (1988).      Other environmental measures  Substrate size composition was assessed using the ‘Wolman Walk’ method where the  β  axis  of  100  stones was measured  at  approximately  1 m  intervals  across  a  zigzag  transect at 45˚ to the stream bank (Wolman 1954).  Percentage substrate composition  of Wentworth scale classes was converted to a single substrate size index by summing  midpoint values of size classes weighted by their proportion.  Bedrock was assigned a  nominal  size  of  400  mm  for  use  in  the  calculations  (Quinn  and  Hickey  1990).  Conductivity,  temperature  and  pH were measured  using  Eutech  instruments  ECScan  pocket  meter  at  New  Zealand  sites  and  a  YSI  556 MPS  meter  at  Spanish  sites.   Depth  and  velocity were  recorded with  a Marsh‐McBirney  flowmate  current  meter  in  the  thalweg of each stream at  five points at equidistant  intervals along  the  study  reach.   Flow  type  of  each  site  was  assessed  visually  as  percentage  of  still,  backwater,  pool,  run  or  riffle  over  a  100 m  reach.   Leaf  litter  percentage,  riparian  vegetation and percent canopy cover were also assessed visually.   Bed  stability/substrate disturbance was  assessed using  the Pfankuch  stability  index  (Pfankuch  1975).   Only  the  bottom  component  of  the  index  (rock  angularity,  brightness,  packing,  percent  stable  materials,  scouring,  and  amount  of  clinging  vegetation) was used, which assesses the substrate component of the stream only, as  Chapter 5: Global productivity‐diversity relationships      131    this  is more  relevant  to  stream  invertebrate  communities  (Winterbourn  and  Collier  1987, Death and Winterbourn 1994).      Statistical analysis  Principal Component Analysis (PCA) was carried out on normalised environmental data  to  determine whether  physicochemical  and  habitat  variable  differed  between  New  Zealand and Spanish streams using Primer v6 (Clarke and Gorley 2006).  Differences in  principal  components  between  regions  were  tested  for  using  one‐way  Analysis  of  Variance  (ANOVA)  in  Statistix  (Statistix  8  ©,  Analytical  Software,  Tallahassee,  FL,  U.S.A.).   Further assessment of whether site characteristics differed between regions  was  assessed  using  Analysis  of  Similarities  (ANOSIM)  (Clarke  1993)  on  normalised  physicochemical data using Euclidean distances in Primer v6 (Clarke and Gorley 2006).   One‐way ANOVA was performed on  individual physicochemical and biological data to  determine whether there were differences in these metrics between Spanish and New  Zealand  streams  using  Statistix  (Statistix  8 ©,  Analytical  Software,  Tallahassee,  FL,  U.S.A.).    Where  required, data were  log  (x + 1)  transformed  in order  to normalise  the  data.    In order  to  test whether diversity was a  log‐linear  function of productivity as  hypothesised,  regression  analysis was  carried  out  to  explore  relationships  between  diversity and chlorophyll a within each region as well as combining both regions and  testing for a universal relationship using Statistix.  Finally, to explore whether stability  was  important  at  determining  diversity  patterns  in  these  streams  regression  was  performed between the Pfankuch index and invertebrate metrics using Statistix, and a  second term was added if it yielded a better fit.  Akaike’s Information Criterion (Akaike  1974) was used to determine the best fitting curve to the data.  This method takes into  account goodness of fit statistics and the number of parameters involved in the fitting  of the model to select the model most preferred.      Chapter 5: Global productivity‐diversity relationships      Results  132    Physicochemical variables  Stream habitat and physicochemical conditions were different between New Zealand  and Spanish  streams  (Fig. 1; Table 1).   Principle  component  (PC) 1  (35.6 % variation  explained)  strongly  split  New  Zealand  and  Spanish  sites  based  on  physicochemical  data, but there was no difference between regions on PC2 (16.4 % variation explained)  (Fig. 1; Table 1).  ANOSIM confirmed there were strong differences in physicochemical  variables between the two regions (R = 0.53, P = 0.001).  Conductivity ranged from 44  to 298 μS cm‐1  in New Zealand and 68 to 402 μS cm‐1  in Spanish streams and was on  average one third higher in the Spanish streams (Table 1).  New Zealand streams were  more acidic, wider, deeper, and had twice the velocity (Fig. 1; Table 1).  There was little  or no difference in slope, substrate size index or substrate heterogeneity between the  two regions.  Bed stability, assessed as the bottom component of the Pfankuch index,  did not differ between New Zealand and Spain (Table 1).      Productivity  Mean  chlorophyll  a  ranged  from  2.58  to  15.35  µg  cm‐2  in  the  Spanish  streams  and  averaged 5.84 µg cm‐2.   Chlorophyll a was on average  three  times  lower  in  the New  Zealand streams ranging from 0.03 to 5.02 µg cm‐2 and averaging 1.88 µg cm‐2 (Table  1).  Chlorophyll a declined with decreasing bed stability at the New Zealand sites (r2 =  0.17,  F  1, 22  = 4.46, P =  0.046,  y = 4.04  ‐  0.06x), however  there was no  relationship  between stability and chlorophyll a at Spanish sites (r2 = 0.06, F 1, 22 = 1.36, P = 0.26).         Figure 1: P the central between  F variation in New Zealan Temp = tem   Cha rincipal com  North Islan ebruary  and  the data (V d, SP = Spa perature, C pter 5: Global productivity‐diversity relationships  ponent ana d of New Ze   July  2007. ectors indic in, SI = subs ond = condu   1 lysis for env aland and C     PC1  expla ate directio trate size ind ctivity.     33  ironmental v antabrian M ined  35.6 % n of increas ex, Subhet  ariables in 2 ountains, n   and  PC2  e e in environ = Shannon s 4 streams f orthern Spa xplained  16 mental varia ubstrate he rom each of in, sampled .4 %  of  the bles).   NZ = terogeneity,             Chapter 5: Global productivity‐diversity relationships      134    Table  1: Mean  (±  1  S.E.)  physicochemical,  periphyton  and  invertebrate  community metrics  collected from 24 streams in each of the central North Island of New Zealand and Cantabrian  Mountains, northern Spain, between February and July 2007.  F and P values indicate results of  one‐way ANOVA  testing  for differences between  the  two  regions.   NZ = New Zealand.   PC =  principle component.                  Site  NZ  Spain  F1, 46 P  Width (m)  12.9 (1.95)  7.72 (1.24)  5.04  0.03  Depth (cm)  31.05 (1.85)  18.86 (0.98)  33.86  0  Velocity (m s‐1)  0.89 (0.04)  0.4 (0.03)  98.48  <0.0001  Conductivity (µs cm‐1)  131.58 (14.15)  205.96 (18.91)  9.92  0.003  Temperature (°C)  10.71 (0.45)  15.15 (0.39)  55.98  <0.0001  pH  7.88 (0.12)  8.86 (0.05)  57.09  <0.0001  Substrate size index  152.59 (9.02)  139.11 (6.08)  1.53  0.222  Substrate heterogeneity  1.9 (0.03)  1.98 (0.03)  3.32  0.075  Slope (m 100 m‐1)  3.77 (0.35)  5.18 (0.61)  4.02  0.051  Pfankuch bottom  36.88 (2.12)  32.75 (1.74)  2.27  0.139  PC1  1.6 (1.19)  ‐1.6 (0.72)  127.57  <0.0001  PC2  0.19 (1.41)  ‐0.19 (1.13)  1.1  0.3  Chlorophyll a (µg cm‐2)  1.88 (0.3)  5.84 (0.72)  26.01  <0.0001  No. of taxa  14.78 (1.02)  15.6 (0.59)  0.48  0.491  No. of animals  327.5 (59.52)  928.23 (406.66)  2.14  0.151  Shannon diversity (H')  1.77 (0.09)  1.63 (0.07)  1.65  0.205  % EPT animals  0.54 (0.06)  0.57 (0.04)  0.24  0.627    Community metrics  We found 76 taxa in Spanish and 82 taxa in New Zealand streams.  The mean number  of  individuals was higher  in  Spanish  than New Zealand  streams but due  to  the high  degree of variation this difference was not significant (Table 1).   Moreover, the mean  number of taxa, Shannon diversity index and the % EPT animals did not differ between  New Zealand and Spanish streams (Table 1).    Density, richness and the Shannon  index  increased  log‐linearly with  increasing  levels of chlorophyll a in the New Zealand streams (Table 2, Fig. 2).  The % EPT animals  was not related to chlorophyll a in either New Zealand or Spanish streams (Table 2, Fig.  2).    The  number  of  animals  increased  linearly with  increasing  chlorophyll  a  in  the  Spanish streams but richness and the Shannon index were not related to chlorophyll a  in these streams (Table 2, Fig. 2).    Chapter 5: Global productivity‐diversity relationships      135      Figure  2: Mean  (±  1  S.E.)  (a,  b)  number  of  animals,  (c,  d)  number  of  taxa,  (e,  f)  Shannon  diversity, and  (g, h) proportion of EPT animals as a  function of chlorophyll a  in 24  (a, c, e, g)  central  North  Island,  New  Zealand  and  (b,  d,  f,  h)  Cantabrian Mountains,  northern  Spain  streams,  collected  between  February  and  July  2007.    Results  of  regression  analysis  are  presented in Table 2.    Table  2:  Results  of  individual  regression  analysis  of  invertebrate  community metrics  as  a  function  of  (a)  chlorophyll  a  and  (b)  the  Pfankuch  bottom  component collected from 24 streams in each of the central North Island of New Zealand and Cantabrian Mountains, northern Spain, between February and  July 2007.   Degrees of  freedom = 1, 22  for  linear and  log‐linear and 2, 21  for quadratic  regressions.   AIC = Akaike’s  information  criterion where  lower  numbers represent a better model.                               New Zealand         Spain               F (AIC)  P  r2 Equation  F  P  r2 Equation  (a) Chlorophyll a (µg cm‐2)  Log10 number of animals  36.07  <0.0001  0.62  y = 2.24 + 0.31∙ln(x)  5.31  0.031  0.19  y = 2.43 + 0.04x  Number of taxa  22.93  <0.0001  0.51  y = 14.41 + 2.79∙ln(x)  0.05  0.83  0.002  Non‐significant  Shannon index  9.87  0.005  0.31  y = 1.75 + 0.18∙ln(x)  0.04  0.83  0.002  Non‐significant  % EPT animals  4.15  0.054  0.16  Non‐significant  1.14  0.3  0.05  Non‐significant  (b) Pfankuch index  Log10 number of animals  12.92 (68.7)  0.002  0.37  y = 3.38 ‐ 0.03x  0.68  0.42  0.03  Non‐significant      quadratic  9.43 (58.6)  0.001  0.47  y = 1.64 + 0.07x ‐ 0.001x2 Number of taxa  4.76 (‐31.9)  0.04  0.18  y = 22.24 ‐ 0.2x  0.19  0.66  0.01  Non‐significant      quadratic  13.42 (‐32.6)  0.0002  0.56  y = ‐10.59 + 1.76x ‐ 0.03x2 Shannon index  1.7 (‐32.2)  0.21  0.07  Non‐significant  0.78  0.39  0.03  Non‐significant      quadratic  7.82 (‐39.2)  0.003  0.43  y = ‐0.48 + 0.15x ‐ 0.002x2 % EPT animals  1.88  0.18  0.08  Non‐significant  0.17  0.69  0.01  Non‐significant    Chapter 5: Global productivity‐diversity relationships      137    Fitting the New Zealand and Spanish invertebrate community metrics together  against chlorophyll a showed that the New Zealand streams fit along the lower end of  the chlorophyll a gradient and  the Spanish streams were at  the higher end  (Table 3,  Fig. 3).  Both the number of animals and the number of taxa fit similar log‐linear curves  to  those  for  the New Zealand data  (Table 3, Fig. 3).   The Spanish data simply placed  further to the right along the curve at higher levels of chlorophyll a (Table 3, Fig. 3).        Figure  3: Mean  (a)  number  of  animals,  (b)  number  of  taxa,  (c)  Shannon  diversity,  and  (d)  proportion  of  EPT  animals  as  a  function  of  chlorophyll  a  in  48  central  North  Island,  New  Zealand and Cantabrian Mountains, northern Spain streams, collected between February and  July 2007.   Black symbols are New Zealand streams and white symbols are Spanish streams.   Results of regression analysis are presented in Table 3.        Chapter 5: Global productivity‐diversity relationships      138    Table  3:  Results  of  combined  regression  analysis  of  invertebrate  community metrics  as  a  function of chlorophyll a collected from 48 streams in the central North Island of New Zealand  and Cantabrian Mountains, northern Spain, between February and July 2007.                      F1, 46 P  r 2 Equation  Log10 number of animals  37.08  <0.0001  0.446  y = 1.27 + 0.11∙ln(x)  Number of taxa  69.88  <0.0001  0.603  y = 2.77 + 0.41∙ln(x)  Shannon index  1.36  0.25  0.029  Non‐significant  % EPT animals  0.02  0.89  0.0004  Non‐significant    None of the four invertebrate community metrics were related to bed stability  in  the Spanish sites  (Table 2).   The number of animals declined at an  increasing  rate  with  stability  in  the New Zealand  streams, and  taxonomic  richness and  the Shannon  diversity  index peaked at  intermediate  levels of stability  in the New Zealand streams  (Table 2).   A  second  term yielded a better  fit  to  the data  for  these metrics  in  these  streams.   There was no relationship between stability and % EPT animals  in the New  .   Zealand streams (Table 2)   Discussion  While  the  productivity‐diversity  relationship  is  a  central  theme  in many  aspects  of  ecology (Abrams 1995, Waide et al. 1999, Mittelbach et al. 2001), until recently  little  emphasis has been placed on assessing this relationship in lotic systems.  Nonetheless,  several recent studies have shown a strong link between the two (Death 2002, Barquin  2004, Death and Zimmermann 2005, Cardinale et al. 2006, Tonkin and Death In prep.‐ a, Tonkin and Death In Prep.‐b, Tonkin et al. In Prep.‐a).  The prevailing theme in these  studies is that the productivity‐diversity relationship is log‐linear at local scales (but see  Death and Zimmermann 2005, Cardinale et al. 2006).  That is, productivity appears to  set  the upper  limit  to  richness  in streams  (Death 2002).   We hypothesized  that both  regions  in  this  study  would  follow  the  same  log‐linear  increase  in  diversity  with  productivity.  At first glance, New Zealand streams were much more closely linked with  productivity  and  exhibited  the  same  log‐linear  relationship  mentioned  above.    In  contrast, no diversity measure appeared  to be  related  to productivity  in  the Spanish  streams.    In  fact,  the only  community metric  linked with productivity  in  the Spanish  Chapter 5: Global productivity‐diversity relationships      139    streams  was  the  number  of  animals  which  increased  linearly  with  productivity.   However, differentiation of relationship between the two regions stream communities  may  simply  be  a  result  of  sampling  at  different  ranges  along  the  productivity  continuum.    Although  it  appears  there  are  two  differing  responses  of  diversity  to  productivity  between  the  Northern  and  Southern  Hemisphere  streams,  Spanish  streams had  roughly  three  times higher primary productivity  than  the New  Zealand  streams.  Essentially, the New Zealand streams examined occurred along the first half  of the productivity scale and Spanish ones along the second half.   Combining the two  regions showed density and richness fit the log‐linear curve fitted to the New Zealand  data, with no evidence of a decline at higher productivity,  suggesting  this may be a  universal pattern.  Most prior studies of this nature in New Zealand streams examined  the  link between productivity  and diversity  at  similar  levels of productivity  to  those  here  (e.g.  Death  2002,  Tonkin  et  al.  In  prep.‐b)  suggesting  these  may  be  simply  covering  the  lower  half  of  the  more  common  unimodal  trend  found  in  other  ecosystems.    However,  Barquin  (2004)  assessed  this  relationship  in  New  Zealand  streams with up to five times higher productivity than the present study and found no  evidence  for a decline  in richness with  increasing productivity.   This  fits with Death’s  (2002)  postulation  that  productivity,  rather  than  competitive  interactions,  sets  the  upper limit to stream richness.  The ‘hump shaped’ relationship between productivity  and diversity is common in many systems, especially with sedentary organisms (Waide  et al. 1999, Mittelbach et al. 2001), and has been found in lotic systems when assessing  primary producer  communities  (Cardinale et  al.  2006).   Rosenzweig  (1995) provides  several  explanations  for  a  decline  in  richness  at  greater  productivity  including  competitive  exclusion,  but  stream  communities  do  not  often  fit within  this  theory.   Nevertheless, competitive exclusion does occur  in stream communities (e.g. Hemphill  and Cooper 1983), but the influence of these interactions appears to be variable (Reice  1985)  and  it  is more  likely  in  streams with  sedentary  species  (Hemphill  and Cooper  1983, McAuliffe 1984).   As most stream  invertebrate communities are dominated by  organisms with  a high degree of mobility  and  immigration  and  emigration  (Downes  Chapter 5: Global productivity‐diversity relationships      140    1990,  Death  2008),  unimodal  productivity‐diversity  relationships  are  rare  (but  see  Death and Zimmermann 2005).    Although the number of taxa  increased with  increasing productivity, evenness  of  communities  declined.    However,  this  was  not  a  consistent  pattern  and  varied  greatly at higher levels of productivity.  Within low productivity streams, Death (2002)  found diversity measures which account  for  community evenness are also positively  linked with productivity.  However, two recent studies which found log‐linear increases  in  richness with productivity  (Tonkin and Death  In prep.‐a, Tonkin et  al.  In prep.‐b),  show that accounting for both evenness and density removed any relationship.  In fact,  combining the New Zealand and Spanish data indicated the response of dominance to  productivity was highly variable.     Tonkin et al.  (In Prep.‐a)  show  that  these Spanish  communities  become  increasingly  dominated  by  Prosimulium  spp.  at  greater  productivity as a result of the high proportion of  filamentous algae  in these streams.   That  is, although black flies (Simuliidae) do not graze directly upon algae they can be  found in high densities feeding on drifting algal cells (Peterson et al. 1985), as well high  densities  of  smaller  individuals  attached  to  filamentous  algae  (Dudley  et  al.  1986).   Tonkin  et  al.  (In  Prep.‐a)  propose  that  within  streams,  the  productivity‐diversity  relationship  is  heavily  influenced  by  the  taxonomic  composition  of  the  algal  communities.  However, although communities became more dominated, competitive  exclusion does not appear to occur.   Several other  lotic studies have also found early  colonisers  remain  after  later  colonisers  arrive  and  are  not  displaced  (Death  1996,  Collier and Quinn 2003, Death 2006).    The %  EPT  index  is  commonly used  as  a bioassessment  tool  in  streams  (e.g.  Lenat  1988,  Barbour  et  al.  1999)  as  these  taxa  are  considered  pollution  sensitive.   However, both within each  region  and  combined,  the  response was highly  variable.   Tonkin et al. (In Prep.‐a) found the percent of EPT animals responds more strongly to  algal  community  composition  than  productivity.    If  eutrophication  only  leads  to  an  increase in productivity, the EPT metric is unlikely to indicate this in our study streams  but if it affects the type of algae it will.  In the New Zealand streams, although richness  increased  along  a  log‐linear  gradient,  there  appeared  to  be  a  strong  shift  in  the  community from EPT dominated to chironomid dominated at roughly the midpoint on  Chapter 5: Global productivity‐diversity relationships      141    the productivity gradient.  However, when compared to the Spanish streams, this shift  in  community  composition  occurred  at  relatively  low  levels  of  productivity which  is  likely  to  be  due  to  the  alpine  nature  of  this  environment.    Specifically,  certain  chironomids are able to tolerate these types of environments such as cold water and  instability and thus appear at lower levels of productivity (Milner and Petts 1994).    Productivity  is  linked  strongly with disturbance  in  streams  (Lake 2000, Death  2002,  Cardinale  et  al.  2006,  Tonkin  and Death  In  prep.‐a,  Tonkin  et  al.  In  prep.‐b).   Consequently, we expected that bed stability could provide explanations for variation  in  the productivity‐diversity relationship.   However,  traditional models  that require a  trade‐off of competitive and colonizing ability  in taxa such as the DEM (Huston 1979,  1994, Kondoh 2001), do not find much support in streams (Tonkin and Death In prep.‐ a, Tonkin et al.  In prep.‐b), but see Cardinale et al.  (2006).   Our  results  indicate  that  New  Zealand  stream  communities  respond  strongly  to  bed  stability  but  Spanish  streams  do  not.    In  fact,  the  Pfankuch  index  was  not  able  to  predict  any  of  the  invertebrate community metrics in Spanish streams and thus stability was not able to  explain the lack of productivity‐diversity relationship in Spanish streams.  Barquin and  Death  (2006)  found  diversity  of  benthic  invertebrates  were  higher  in more  stable  habitats  in New Zealand streams but Barquin and Death (2004) found the opposite  in  Spanish  streams.   New  Zealand  stream  communities  are  highly  dependent  on  both  productivity and disturbance for the maintenance of diversity (Death 2002, Tonkin et  al. In prep.‐b), but these studies suggest the role of disturbance  is simply to reset the  colonization process (at different rates) allowing habitats to be recolonised.   Richness  is consequently governed by productivity through the processes discussed previously;  however this was not exhibited in the Spanish streams which may be due to a lack of  low productivity sites in the limited range of study sites.    The  lack  of  focused  assessments  of  the  productivity‐diversity  relationship  in  lotic  systems  has  left  benthic  ecologists  with  no  clear  idea  of  the  nature  of  the  relationship.    With  the  strong  role  productivity  plays  in  streams,  especially  in  conjunction with disturbance  (Lake  2000, Death  2002, Cardinale  et  al.  2006,  Tonkin  and  Death  In  prep.‐a,  Tonkin  et  al.  In  prep.‐b),  it  is  essential  to  elucidate  this  relationship.   As has been found  in recent studies (Death 2002, Barquin 2004, Tonkin  Chapter 5: Global productivity‐diversity relationships      142    and Death  In prep.‐a, Tonkin et al.  In prep.‐b), we suggest  the  role of productivity  is  indeed  to dictate maximum  richness  in streams  rather  than  to  increase  the effect of  competitive interactions.  Although variation in response between the two regions was  evident  in  this  study,  the overall  relationship when  combining  regions  suggests  that  Death’s (2002) postulation of productivity setting the upper limit to richness in streams  is  correct.    The  variable  response  found  between  the  two  regions  indicates  that  caution  is  required when  interpreting  results along differing  scales of environmental  variation.  With the large range of productivity assessed in this study and no decline in  richness  at  high  levels  of  productivity, we  provide  strong  evidence  that  the DEM  is  treams.   unlikely to be relevant in s   References  Abrams,  P. A.  1995. Monotonic  or  unimodal  diversity  productivity  gradients  ‐ What  does competition theory predict. Ecology 76:2019‐2027.  Akaike,  H.  1974.  New  look  at  statistical‐model  identification.  Ieee  Transactions  on  Automatic Control AC19:716‐723.  Barbour, M. T., J. Gerritsen, B. D. Snyder, and J. B. Stribling. 1999. Rapid Bioassessment  Protocols  for  Use  in  Streams  and  Wadeable  Rivers:  Periphyton,  Benthic  Macroinvertebrates,  and  Fish.  Second  Edition.  EPA  841‐B‐99‐002.  U.S.  Environmental Protection Agency, Office of Water, Washington, D.C.  Barquin, J. 2004. Spatial patterns of invertebrate communities in spring and runoff‐fed  streams.  PhD  thesis.  PhD  thesis, Massey  University,  Palmerston  North,  New  Zealand.  Barquin,  J., and R. G. Death. 2004. Patterns of  invertebrate diversity  in  streams and  freshwater springs in Northern Spain. Archiv Fur Hydrobiologie 161:329‐349.  Barquin,  J., and R. G. Death. 2006. Spatial patterns of macroinvertebrate diversity  in  New Zealand springbrooks and rhithral streams. Journal of the North American  Benthological Society 25:768‐786.  Chapter 5: Global productivity‐diversity relationships      143    Cardinale, B. J., H. Hillebrand, and D. F. Charles. 2006. Geographic patterns of diversity  in  streams  are  predicted  by  a  multivariate  model  of  disturbance  and  productivity. Journal of Ecology 94:609‐618.  Cardinale, B. J., M. A. Palmer, A. R. Ives, and S. S. Brooks. 2005. Diversity‐productivity  relationships  in streams vary as a  function of  the natural disturbance  regime.  Ecology 86:716‐726.  Chase,  J.  M.,  and  M.  A.  Leibold.  2002.  Spatial  scale  dictates  the  productivity‐ biodiversity relationship. Nature 416:427‐430.  Clarke,  K.  R.  1993.  Nonparametric  multivariate  analyses  of  changes  in  community  structure. Australian Journal of Ecology 18:117‐143.  Clarke,  K.  R.,  and  R. N. Gorley.  2006.  PRIMER  v6: User Manual/Tutorial.  PRIMER‐E,  Plymouth.  Collier, K. J., and J. M. Quinn. 2003. Land‐use influences macroinvertebrate community  response following a pulse disturbance. Freshwater Biology 48:1462‐1481.  Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs. Science 199:1302‐ 1310.  Currie, D. J. 1991. Energy and large‐scale patterns of animal‐species and plant‐species  richness. American Naturalist 137:27‐49.  Currie,  D.  J.,  and  V.  Paquin.  1987.  Large‐scale  biogeographical  patterns  of  species  richness of trees. Nature 329:326‐327.  Death, R. G. 1996. The effect of patch disturbance on stream invertebrate community  structure: the influence of disturbance history. Oecologia 108:567‐576.  Death, R. G. 2002. Predicting invertebrate diversity from disturbance regimes in forest  streams. Oikos 97:18‐30.  Death,  R.  G.  2006.  Colonisation  in  New  Zealand  streams:  predictable  patterns  or  chance events? New Zealand Natural Sciences 31:93‐112.  Death, R. G. 2008. Effects of floods on aquatic invertebrate communities.in J. Lancaster  and R. A. Briers, editors. Insects: Challenges to Populations. CAB International,  UK, pp 103‐121.  Chapter 5: Global productivity‐diversity relationships      144    Death,  R. G.,  and M.  J. Winterbourn.  1994.  Environmental  stability  and  community  persistence:  a  multivariate  perspective.  Journal  of  the  North  American  Benthological Society 13:125‐139.  Death,  R.  G.,  and  M.  J.  Winterbourn.  1995.  Diversity  patterns  in  stream  benthic  invertebrate communities:  the  influence of habitat  stability. Ecology 76:1446‐ 1460.  Death,  R.  G.,  and  E. M.  Zimmermann.  2005.  Interaction  between  disturbance  and  primary  productivity  in  determining  stream  invertebrate  diversity.  Oikos  111:392‐402.  Downes,  B.  J.  1990.  Patch  dynamics  and  mobility  of  fauna  in  streams  and  other  habitats. Oikos 59:411‐413.  Dudley, T. L., S. D. Cooper, and N. Hemphill. 1986. Effects of Macroalgae on a Stream  Invertebrate Community. Journal of the North American Benthological Society  5:93‐106.  Gaston, K. J. 2000. Global patterns in biodiversity. Nature 405:220‐227.  Graham, A. A., D. J. McCaughan, and F. S. McKee. 1988. Measurement of surface area  of stones. Hydrobiologia 157:85‐87.  Grime,  J. P. 1973a. Competitive exclusion  in herbaceous vegetation. Nature 242:344‐ 347.  Grime,  J.  P.  1973b.  Control  of  species  density  in  herbaceous  vegation.  Journal  of  Environmental Management 1:151‐167.  Hemphill, N., and S. D. Cooper. 1983. The effect of physical disturbance on the relative  abundances of 2 filter‐feeding insects in a small stream. Oecologia 58:378‐382.  Hubbell,  S.  P.  2001.  The  unified  neutral  theory  of  biodiversity  and  biogeography.  Princeton University Press.  Huston, M. 1979. A general hypothesis of  species diversity. The American Naturalist  113:81‐100.  Huston,  M.  1994.  Biological  Diversity:  The  Coexistence  of  Species  on  Changing  Landscapes. Cambridge University Press, Cambridge, UK.  Kneitel, J. M., and J. M. Chase. 2004. Disturbance, predator, and resource interactions  alter container community composition. Ecology 85:2088‐2093.  Chapter 5: Global productivity‐diversity relationships      145    Kondoh, M.  2001. Unifying  the  relationships  of  species  richness  to  productivity  and  disturbance.  Proceedings  of  the  Royal  Society  of  London  Series  B‐Biological  Sciences 268:269‐271.  Lake, P. S. 2000. Disturbance, patchiness, and diversity in streams. Journal of the North  American Benthological Society 19:573‐592.  Leibold,  M.  A.  1999.  Biodiversity  and  nutrient  enrichment  in  pond  plankton  communities. Evolutionary Ecology Research 1:73‐95.  Lenat, D. R. 1988. Water quality assessment of streams using a qualitative collection  method for benthic invertebrates. Journal of the North American Benthological  Society 7:222‐233.  Mackey,  R.  L.,  and  D.  J.  Currie.  2001.  The  diversity‐disturbance  relationship:  Is  it  generally strong and peaked? Ecology 82:3479‐3492.  McAuliffe,  J.  R.  1984.  Competition  for  space,  disturbance,  and  the  structure  of  a  benthic stream community. Ecology 65:894‐908.  Milner,  A. M.,  and  G.  E.  Petts.  1994.  Glacial  rivers  ‐  physical  habitat  and  ecology.  Freshwater Biology 32:295‐307.  Mittelbach, G. G., C. F. Steiner, S. M. Scheiner, K. L. Gross, H. L. Reynolds, R. B. Waide,  M. R. Willig, S. I. Dodson, and L. Gough. 2001. What is the observed relationship  between species richness and productivity? Ecology 82:2381‐2396.  Morin, A., W. Lamourex, and  J. Busnarda. 1999. Empirical models predicting primary  productivity  from chlorophyll a and water temperature  for stream periphyton  and  lake  and  ocean  phytoplankton.  Journal  of  the  North  American  Benthological Society 18:299‐307.  Palmer,  M.  W.  1994.  Variation  in  species  richness  ‐  towards  a  unification  of  hypotheses. Folia Geobotanica & Phytotaxonomica 29:511‐530.  Peterson, B.  J.,  J. E. Hobbie, A. E. Hershey, M. A.  Lock, T. E.  Ford,  J. R. Vestal, V.  L.  McKinley, M. A.  J. Hullar, M. C. Miller, R. M. Ventullo,  and G.  S. Volk.  1985.  Transformation of a Tundra River from Heterotrophy to Autotrophy by Addition  of Phosphorus. Science 229:1383‐1386.  Pfankuch, D. J. 1975. Stream Reach  Inventory and Channel Stability Evaluation. USDA  Forest Service, Region 1, Missoula, Montana.  Chapter 5: Global productivity‐diversity relationships      146    Pickett, S. T. A., and P. S. White. 1985. The ecology of natural disturbance and patch  dynamics. Academic Press.  Reice, S. R. 1985. Experimental disturbance and the maintenance of species‐diversity in  a stream community. Oecologia 67:90‐97.  Resh, V. H., A. V. Brown, A. P. Covich, M. E. Gurtz, H. W. Li, G. W. Minshall, S. R. Reice,  A. L. Sheldon, J. B. Wallace, and R. C. Wissmar. 1988. The role of disturbance in  stream  ecology.  Journal  of  the North  American  Benthological  Society  7:433‐ 455.  Ricklefs, R.  E.  2004. A  comprehensive  framework  for  global patterns  in biodiversity.  Ecology Letters 7:1‐15.  Rosenzweig, M. L. 1992. Species‐diversity gradients ‐ we know more and less than we  thought. Journal of Mammalogy 73:715‐730.  Rosenzweig, M.  L.  1995.  Species Diversity  in  Space  and  Time.  Cambridge University  Press, Cambridge, UK.  Rosenzweig, M. L., and Z. Abramsky. 1993. How are Diversity and Productivity Related?  Pages  52‐65  in  R.  E.  Ricklefs  and  D.  Schluter,  editors.  Species  Diversity  in  Biological Communities. University of Chicago Press, Chicago, Illinois, USA.  Scholes, L., P. H. Warren, and A. P. Beckerman. 2005. The combined effects of energy  and  disturbance  on  species  richness  in  protist  microcosms.  Ecology  Letters  8:730‐738.  Sousa, W. P. 1979. Disturbance in marine intertidal boulder fields: the nonequilibrium  maintenance of species diversity. Ecology 60:1225‐1239.  Sousa, W. P. 1984. The role of disturbance  in natural communities. Annual Review of  Ecology and Systematics 15:353‐391.  Steinman, A. D.,  and G. A.  Lamberti.  1996. Biomass  and  pigments  of  benthic  algae.  Pages 295‐314  in F. R. Hauer and G. A. Lamberti, editors. Methods  in Stream  Ecology. Academic Press, San Diego, CA.  Tachet, H., P. Richoux, M. Bournaud, and P. Usseglio‐Polatera. 2000. Invertebres d'eau  douce. Systematique, biologie, ecologie. CNRS Editions, Paris.  Tonkin,  J.  D.,  and  R.  G.  Death.  In  prep.‐a.  Consistent  effects  of  productivity  and  disturbance on diversity between landscapes.  Chapter 5: Global productivity‐diversity relationships      147    Tonkin, J. D., and R. G. Death.  In Prep.‐b. Scale dependent effects of productivity and  disturbance on diversity.  Tonkin, J. D., R. G. Death, and J. Barquin. In Prep.‐a. Productivity‐diversity relationships  in streams and the role of periphyton assemblages.  Tonkin, J. D., R. G. Death, and K. J. Collier. In prep.‐b. Do productivity and disturbance  interact to modulate macroinvertebrate diversity in streams?  Towns,  D.  R.,  and  W.  L.  Peters.  1996.  Leptophlebiidae  (Insecta:  Ephemeroptera).  Manaaki Whenua Press, Lincoln, New Zealand.  Townsend, C. R., M. R. Scarsbrook, and S. Doledec. 1997. The intermediate disturbance  hypothesis, refugia, and biodiversity  in streams. Limnology and Oceanography  42:938‐949.  Vinson, M. R., and C. P. Hawkins. 1998. Biodiversity of stream insects: variation at local,  basin, and regional scales. Annual Review of Entomology 43:271‐293.  Waide, R. B., M. R. Willig, C.  F.  Steiner, G. Mittelbach,  L. Gough,  S.  I. Dodson, G. P.  Juday,  and  R.  Parmenter.  1999.  The  relationship  between  productivity  and  species richness. Annual Review of Ecology and Systematics 30:257‐300.  Wilson,  S. D.,  and D.  Tilman.  2002. Quadratic  variation  in  old‐field  species  richness  along gradients of disturbance and nitrogen. Ecology 83:492‐504.  Winterbourn, M. J., and K. J. Collier. 1987. Distribution of benthic invertebrates in acid,  brown  water  streams  in  the  South  Island  of  New  Zealand.  Hydrobiologia  153:277‐286.  Winterbourn, M.  J., K.  L. D. Gregson, and C. H. Dolphin. 2000. Guide  to  the aquatic  insects of New Zealand. Entomological Society of New Zealand, Auckland.  Wolman, M. J. 1954. A method of sampling coarse river bed material. Transactions of  the American Geophysical Union 35:951‐956.  Wootton,  J.  T.  1998.  Effects  of  disturbance  on  species  diversity:  A  multitrophic  perspective. American Naturalist 152:803‐825.          Chapter 5: Global productivity‐diversity relationships      148    Appendix 1: Latitude, longitude and altitude of 48 streams sampled in the central North Island of New Zealand and Cantabrian Mountains of northern Spain,  between February and July 2007.             New Zealand    Spain    Site  Longitude Latitude Alt. Site Longitude Latitude Alt.  Mangatoetoenui @ SH1 175.7316156 ‐39.2363068 971 Rio Besaya @ Helguera ‐4.03276 43.158628 214  Whakapapa d/s intake 175.4659706 ‐39.11992199 680 Rio Eracia @ Helguera ‐4.030378 43.159622 222  Whakapapaiti @ SH4 175.4720796 ‐39.17692963 859 Rio Bisuena @ Barcena Pie de Concha ‐4.06826 43.121292 309  Trib of Ohinepango @ old hut  175.671968 ‐39.20895258 1106 Torina @ Barcena Pie de Concha ‐4.054722 43.126793 299  Waihohonu Springs 175.6654827 ‐39.18896035 1158 Santiurde @ Santiurde ‐4.078611 43.063231 634  Wahianoa stream u/s intake  175.6407765 ‐39.38396292 934 Rumadero @ Pesquera ‐4.074784 43.083177 574  Te Unuunuakapuateariki stream  175.5593075 ‐39.4431139 701 Rio Leon @ San Martin de Quevedo ‐4.038779 43.139996 264  Tongariro d/s Poutu intake  175.8224059 ‐39.13002278 456 Ayo Valdeiguna @ Pedredo ‐4.067773 43.192911 185  Waipakihi @ end of road 175.7729642 ‐39.22440144 857 Rio Argoza @ Barcena Mayor ‐4.232753 43.156963 422  Mangatepopo by camp 175.5633283 ‐39.06976046 752 Rio Saja @ Renedo ‐4.30455 43.194076 293  Mangatepopo d/s dam 175.5523842 ‐39.05335106 624 Rio de la Magdalena @ San Andres ‐3.897027 43.096378 412  Ohinepango Springs 175.6779304 ‐39.21838435 1091 Ayo Salcera @ San Miguel de Luena ‐3.899921 43.112736 347  Mangaturuturu river 175.3917215 ‐39.30696099 820 Rio Viaña @ Viaña ‐3.804674 43.155633 326  Tongariro d/s Rangipo dam  175.779722 ‐39.21008569 813 Rio Pas @ Pandillo ‐3.759107 43.164859 366  Tongariro @ pillars of hercules  175.7881839 ‐39.18527209 660 Llerana @ Coterillo ‐3.794589 43.263832 207  Whanganui @ Te Porere 175.5924007 ‐39.04622437 654 Rio Pisuena @ Barcena de Carriedo ‐3.823485 43.241998 181  Whanganui ds intake 175.6010392 ‐39.01997827 604 Nansa @ Puentenansa ‐4.406832 43.257111 168  Oturere Stream SH1 175.7576087 ‐39.18382654 809 Ayo de Hoyamala @ Puentenansa ‐4.406179 43.257101 163  Unnamed @ SH1 175.7578776 ‐39.1940232 885 Lamasón @ Quintanilla ‐4.476253 43.25448 238  Whakapapanui @SH4 175.5090823 ‐39.1491875 835 Ayo de Traveseras @ Quitanilla ‐4.30588 43.256427 239  Unnamed Karioi forest stream  175.6278138 ‐39.38270332 935 Palomba @ Paracuelles ‐4.211634 43.019856 895  Poutu Stream 175.7969903 ‐39.07827181 518 Rio Hijar @ Espanilla ‐4.226527 43.019571 937  Tauwhitikuri Stream d/s intake  175.5513666 ‐39.05365336 621 Rio Rucebos @ Soto ‐4.222249 43.035916 960  Makomiko stream 175.3923096 ‐39.23638244 746 Rio Guares @ Abiado ‐4.289066 43.016011 1061    Chapter 6  Scale dependent effects of productivity and  disturbance on diversity                      Chapter 6: Scale dependent links between productivity, disturbance and diversity      Abstract  150    The effects of productivity and disturbance on diversity are highly  variable with  the  spatial  scale at which  they are examined.   Not only do productivity and disturbance  have strong  influences on diversity patterns at  local and regional scales but they can  affect  the way  in which  communities  assemble  and  in  turn  alter  beta  diversity  or  community dissimilarity.  We assessed whether the form of both the productivity‐ and  disturbance‐diversity  relationships  differed  between  the  spatial  scale  at which  they  were  examined  using  experimental  stream  channels  in  three  Hawke’s  Bay,  New  Zealand,  streams.    In place of  true  local  and  regional  richness, we used  the proxies  within‐ and between‐stream richness, as well as assessing between‐stream community  dissimilarity  (similar  to  beta  diversity).    Our  results  indicate  that  productivity  and  disturbance both affect diversity but at individual scales and in different forms.  At the  within‐stream scale, richness was a u‐shaped  function of productivity whereas at the  between‐stream  scale  richness  increased monotonically with  increasing productivity.   Community dissimilarity on  the other hand,  increased monotonically with  increasing  rate  of  disturbance.    Rather  than  a  greater  role  of  deterministic  assembly  with  increasing  disturbance,  our  results  indicate  the  opposite  but  it  appears  that  communities are simply converging on those found in the surrounding streambed with  time since disturbance.   Specifically, communities were more similar within individual  streams  than within  disturbance  treatments  and  those  to  colonise  post‐disturbance  were  simply a  subset of  taxa present at each  site  regardless of perceived  colonising  ability rather than a suite of specialist colonising taxa.  These results demonstrate that  without a distinction between early and late colonisers, a greater rate of deterministic  assembly at high disturbance will not occur.      Chapter 6: Scale dependent links between productivity, disturbance and diversity      Introduction  151    The  relationship  between  productivity  and  diversity  can  be  highly  variable  (Currie  1991, Abrams 1995, Waide et al. 1999, Mittelbach et al. 2001).  Unimodal (e.g. Grime  1973a,  Huston  1979,  Rosenzweig  1992,  1995,  Leibold  1999,  Waide  et  al.  1999,  Mittelbach et  al.  2001) or  linear  (e.g. Currie  and Paquin  1987, Currie  1991, Abrams  1995, Waide et al. 1999, Gaston 2000, Mittelbach et al. 2001)  increases  in diversity  with increasing productivity are those most often observed.  However, linear declines,  u‐shaped relationships and no relationship also occur (Waide et al. 1999, Mittelbach et  al. 2001).   One explanation for the different outcomes  is the difference  in the spatial  scale  of  sampling  (Currie  1991,  Chase  and  Leibold  2002).    Smaller  scale  studies  commonly  find  unimodal  relationships  between  productivity  and  diversity  but  at  increasing  spatial  scales  the pattern  is often one of monotonic  increases  in diversity  with  productivity  (Waide  et  al.  1999,  Mittelbach  et  al.  2001).    There  are  many  mechanisms  postulated  to  lead  to  this  unimodal  relationship  (Rosenzweig  and  Abramsky 1993, Abrams 1995, Waide et al. 1999).   For example, Kassen et al. (2000)  have  shown  niche  specialisation  in  heterogeneous,  but  not  homogeneous  environments can cause this pattern.  Chase and Leibold (2002) also found differences  in  the  nature  of  the  relationship  when  considered  from  different  scales;  a  hump  shaped  trend  occurred  at  local  scales,  a  linear  pattern  at  the  regional  scale  and  community dissimilarity increased with productivity.    The  mechanisms  underlying  the  differences  in  the  relationship  between  productivity and diversity  remain uncertain, although many have been hypothesised  and these are also likely to vary with scale (Waide et al. 1999).  Two factors thought to  strongly influence the productivity‐diversity relationship are disturbance (Huston 1979,  1994,  Kondoh  2001),  and  the  history  of  community  assembly  (Fukami  and Morin  2003).    The  disturbance‐diversity  relationship  also  varies with  spatial  and  temporal  scale  (Petraitis  et  al.  1989, Mackey  and  Currie  2001,  Chase  2007).    However  the  prevailing theme in disturbance ecology revolves around the Intermediate Disturbance  Hypothesis (IDH) (Grime 1973b, Connell 1978, Sousa 1979), and recently Roxburgh et  al. (2004) have shown the promotion of richness at intermediate levels of disturbance  Chapter 6: Scale dependent links between productivity, disturbance and diversity      152    can  come  from  many  mechanisms.    But  the  requirement  for  a  competition‐ colonisation trade‐off (Chesson and Huntly 1997, Roxburgh et al. 2004, Cadotte 2007),  and the fact that this has been found to be lacking in many ecosystems, has led many  to  suggest  disturbance  simply  removes  taxa  (Death  and Winterbourn  1995).   Many  have advocated that diversity is a function of the interaction between disturbance and  productivity (Huston 1979, 1994, Kondoh 2001, Cardinale et al. 2006), however recent  empirical work suggests the effects are additive (Tonkin and Death In prep., Tonkin et  al. In prep.).    Both  productivity  and  disturbance  can  also  alter  β  diversity  and  community  dissimilarity  through  changes  in  assembly  sequences  (Chase  2007,  Lepori  and  Malmqvist  2009,  Chase  2010).    Deterministic  assembly  involves  the  recruitment  of  colonists  from  the  regional  pool  based  on  niche  preferences,  whereas  stochastic  assembly  involves  random  selection of  available  colonists  (i.e. ecological drift).    The  debate on whether deterministic  (Poff 1997) or stochastic determinates  (Reice 1994,  Hubbell  2001)  are  more  important  is  long  standing,  though  it  is  likely  that  a  combination  of  both  occur  (Hart  1992,  Thompson  and  Townsend  2006).    Recently,  Chase (2010) suggested that more productive environments lead to higher β diversity  through a stronger dependence on stochastic, as opposed to deterministic, community  assembly.    Deterministic  assembly  processes will  likely  be more  prevalent  in  harsh  environments where conditions filter out unsuitable taxa; benign environments on the  other  hand  will  be  governed  by more  stochastic  processes  allowing  for  a  greater  representation  of  the  regional  pool  (Chase  2007).    This  suggests  β  diversity  or  community dissimilarity will decline with disturbance rate as a  limited range of more  specialist  traits  are  required  in more  disturbed  habitats.    Nonetheless,  Lepori  and  Malmqvist (2009) found deterministic control (and associated  lowest β diversity) was  greatest  at  intermediate  levels  of  disturbance  suggesting  a more  complex  interplay  between stochastic and deterministic control.    We set out to test the effect of spatial scale on both the productivity‐diversity  and  disturbance‐diversity  relationships  using  experimental  stream  channels.    Rather  than true α, β and γ diversity, we used within‐  (local) and between‐stream (regional)  richness  and  between  stream  community  dissimilarity  (β  diversity)  for  each  Chapter 6: Scale dependent links between productivity, disturbance and diversity      153    disturbance  and  productivity  treatment.    Based  on  previous  work  suggesting  productivity sets the upper limit to diversity in streams (Death 2002, Tonkin and Death  In prep., Tonkin et al. In prep.), we predict that within‐stream taxonomic richness will  increase  log‐linearly  with  productivity  and  between‐stream  richness  will  increase  monotonically (Chase and Leibold 2002).  As there is no clear competitive‐colonisation  trade‐off  in New Zealand  stream  invertebrate  communities  (Death and Winterbourn  1995,  Death  2002),  we  hypothesise  that  within‐stream  richness  will  decline  with  disturbance.  Although there is support for increased diversity with disturbance at the  regional  scale,  as  it  allows  for  the  representation  of  early  colonising  species within  patchy  environments  (Hastings  1980,  Pickett  and White  1985,  Chesson  and  Huntly  1997), we predict between‐stream richness should match within‐stream richness and  decline.    This  is  because  between‐stream  richness  was  assessed  at  the  same  disturbance rate, thus disturbance should theoretically be homogeneous.   We expect  greater  dissimilarity  in  community  composition  (β  diversity)  at  lower  rates  of  disturbance as a result of more stochastic assembly processes (Chase 2003, 2007).  The  scale of this study  is such that connectivity between replicates  is  low which has been  found to confound diversity assessments (Warren 1996, Chase 2003, Chase and Ryberg  006, Matthiessen et al. 2010).   2004, Ostman et al. 2   Methods  Study sites and physicochemical measures  Pastoral farming and wine production dominates land‐use of Hawke’s Bay, in the East  of the North Island of New Zealand; an area characterised by a warm and dry climate  with a mean annual rainfall of 783 mm.  Three spring‐fed streams of relatively similar  characteristics (mean width: 1.5‐4.7 m; mean depth: 0.15‐0.44 m; mean velocity: 0.42‐ 1.52 m s‐1) were selected for use in this study in the Ruataniwha Plains.  Temperature  and conductivity in these streams ranged from 15.2 to 17.9˚C and 160 to 220 µS cm‐1  respectively during the experimental period and streams range in altitude from 134 to  146 m a.s.l..     Chapter 6: Scale dependent links between productivity, disturbance and diversity      154    Experimental methods  Eight plastic channels (1500 mm long x 150 mm wide x 100 mm deep) were placed in  the bed of each stream and filled with cobbles.  Cobbles were sourced from within the  stream bed and allowed to acclimatise for 21 days before the treatment period.  Four  disturbance and  two productivity  treatments were applied within each  stream.   The  disturbance treatments were: no disturbance (after initial disturbance at day 0), every  16, 8, and 4 days; and productivity treatments were either high or  low.     This yielded  eight  treatments:  the  four  levels  of  disturbance  each  at  low  and  high  productivity  levels.  Productivity was characterised as high or low based on the presence or absence  of shading.   Shade was applied by covering  four of the channels  in each stream with  1800 mm x 350 mm metal sheeting approximately 200 mm above  the substrate and  held in place with metal stakes and rubber grommets.  Disturbance was performed by  physically  stirring  the  full  contents  of  the  channel  whilst  minimising  removal  of  periphyton on  the  substrate  for  two minutes.   However,  larger  forms of macroalgae  were inevitably removed, if present, due to their ease of detachment.      Sampling protocols  Invertebrates were sampled from three random 250 mm sections within each channel.   Sampling was performed with a modified section of guttering, similar to the artificial  channels, with a 250 µm mesh net attached.  One stone (α axis < 60 mm) was removed  from each  replicate  for  later analysis of periphyton biomass.   The  remaining  sample  was placed in 70 % ETOH and later identified in the laboratory using available keys (e.g.  Towns and Peters 1996, Winterbourn et al. 2000).    Primary productivity was estimated from measures of chlorophyll a on stones  from within each sample.  Chlorophyll a of natural substrate and primary productivity  in streams are highly correlated  (Morin et al. 1999, Tonkin and Death In prep.).  Stones  were  kept  cool  and  dark  on  ice  in  the  field  before  being  stored  at  ‐20˚C.   Photosynthetic  pigments  were  extracted  from  stones  by  submerging  in  known  volumes of 90 % acetone  for 24 hours at 5˚C.   Absorbances at 750, 665 and 664 nm  were  read on a Varian Cary 50  conc UV‐Visible Spectrophotometer  (Varian Australia  Pty  Ltd, Mulgrave, Australia)  and  converted  to  chlorophyll  a  pigment  concentration  Chapter 6: Scale dependent links between productivity, disturbance and diversity      155    using  Steinman  and  Lamberti  (1996).    These were  then  corrected  for  stone  surface  area (Graham et al. 1988) and halved to account for periphyton being present only on  upper stone surfaces.      Statistical analysis  Within‐stream  (‘local’)  taxonomic  richness was  calculated  as  the  pooled  number  of  taxa  for  the  three  samples  within  each  treatment.    Between‐stream  (‘regional’)  richness was calculated as the total number of taxa for the three samples within each  treatment and each stream (nine samples).  This is not true regional richness (γ) but a  surrogate  as  it  is  simply  the  total  number  of  taxa  in  all  three  streams  for  each  treatment,  thus  we  use  the  terms  within‐  and  between‐stream  richness.    We  calculated the dissimilarity in community composition in treatments between streams.   These  dissimilarity  metrics  left  us  with  three  pairwise  comparisons  which  were  averaged to achieve a ‘regional’ dissimilarity metric.  Dissimilarity (100 – similarity) was  estimated using the Bray‐Curtis similarity metric in Primer v6 (Clarke and Gorley 2006).   This method takes into account variation in taxonomic abundance and was performed  on raw data.  Although this metric is similar conceptually to β diversity, it is a measure  of  community  dissimilarity  accounting  for  abundances  rather  than  the  number  of  unique taxa to each community.  Thus, if dissimilarity is zero, all taxa are shared and at  the same abundances between all communities and if dissimilarity is 100 then no taxa  are  shared.    In  order  to  assess  whether  changes  in  taxonomic  richness  with  productivity and disturbance were due  to changes  in  the density of animals, we also  calculated the number of individuals and rarefied taxonomic richness.    To  assess  one  factor  of  resource  heterogeneity  we  calculated  variation  in  productivity  between  replicates  and  streams  by  summing  the  differences  between  chlorophyll a values with each treatment.  Similarity Percentages (SIMPER) was used to  assess  which  taxa  contributed  the  greatest  to  differences  between  and  within  treatments using Primer v6.   Linear and quadratic regression was performed  in order  to  test  for  relationships between productivity, disturbance and diversity at  the  three  spatial scales using Statistix (Statistix 8 ©, Analytical Software, Tallahassee, FL, U.S.A.).   The  effects  of  productivity  and  stability  based  on  a  priori  group  selection  were  Chapter 6: Scale dependent links between productivity, disturbance and diversity      156    assessed using two‐way crossed Analysis of Variance (ANOVA) design in Statistix.  Both  tability were treated as fixed factors.   productivity and s   Results  Chlorophyll a was higher  in all high productivity  treatments  (F1, 22 = 4.36, P = 0.049),  but did not differ with disturbance rate (F3, 20 = 0.75, P = 0.54).  The number of taxa did  not differ with disturbance rate (F3, 16 = 0.33, P = 0.81), productivity (F1, 16 = 0.15, P =  0.7), or an interaction between the two (F3, 16 = 0.13, P = 0.93).    Within‐stream taxonomic richness exhibited a u‐shaped quadratic response to  increasing  productivity  (Fig.  1,  Table  1).    This  u‐shaped  relationship  between  productivity  and  richness  at  the  local  scale  can  be  explained  by  two  separate  relationships  related  to  the  productivity  treatment.    In  the  low  productivity  treatments,  richness declined  at  a decreasing  rate with  increasing productivity  (r2  =  0.68,  F2, 9  =  9.53,  P  =  0.01,  y  =  15.2  –  20.82x  +  18.62x 2).    In  the  high  productivity  treatments, richness increased monotonically although only at the 1 % level (r2 = 0.32,  F1,  10  =  4.74,  P  =  0.06,  y  =  8.84  +  3.58x).    Richness  declined  monotonically  with  productivity  at medium  disturbance  sites  (8  day  disturbance  rate)  (r2  =  0.94,  F1, 4  =  59.81, P = 0.002, y = 13.38 – 6.2x) but was not related at any other disturbance rate.      Table  1:  Results  of  linear  and  quadratic  regression  analysis  for  (a)  productivity‐  and  (b)  disturbance‐diversity  relationships  at  within‐  and  between‐stream  scales  and  community  dissimilarity (100 – Bray‐Curtis dissimilarity) in artificial channels in Hawke’s Bay, New Zealand  streams, December 2009.       F  P  r2 Equation  (a) Productivity  Within‐stream richness  4.98  0.02  0.32  y = 13.25 ‐ 23.27x + 11.49x2 Between‐stream richness  16.68  0.007  0.74  y = 14.42 + 8.76x  Community dissimilarity  0.025  0.88  0.004  y = 66.13 + 4.46x  (b) Disturbance  Within‐stream richness  0.01  0.92  0.0004  y = 10.75 + 0.03x  Between‐stream richness  0.028  0.87  0.005  y = 18.5 + 0.1x  Community dissimilarity  10.16  0.02  0.63  y = 48.71 + 7.85x    Chapter 6: Scale dependent links between productivity, disturbance and diversity      157      Figure 1: Linear and quadratic regression between diversity and (a, c, e) productivity and (b, d,  f)  disturbance  at  (a,  b)  within‐  and  (c,  d)  between‐stream  scales  and  (e,  f)  community  dissimilarity (100 – Bray‐Curtis dissimilarity) in artificial channels in Hawke’s Bay, New Zealand  streams, December  2009.   Community dissimilary points  represent  the  average of  two‐way  comparisons for each treatment between streams.  Open symbols represent high productivity  and closed low productivity treatments.  Upright triangles represent disturbance rate of 1 (not  disturbed), circles 2 (every 16 days), squares 3 (every 8 days), and inverted triangles 4 (every 4  days).          Chapter 6: Scale dependent links between productivity, disturbance and diversity      158    Between‐stream richness  increased monotonically with  increasing productivity  (Fig. 1, Table 1).   However,  community dissimilarity was not  related  to productivity.   The  average  variation  in  productivity  (a  surrogate  for  resource  heterogeneity)  increased with  increasing productivity but was not significant  (F1, 6 = 3.33, P = 0.12).   Rarefied richness was not related  to productivity at  this scale  (F1, 6 = 0.05, P = 0.83),  but the number of animals also increased with productivity at the 1 % level (r2 = 0.44,  F1, 6 = 4.71, P = 0.07, y = 28.8 + 108.38x).    Within‐ and between‐stream richness were not related to disturbance rate but  community dissimilarity  increased monotonically with  increasing  rate of disturbance  (Fig.  1,  Table  1).    The  number  of  individuals  increased with  disturbance  rate  at  the  between‐stream  level (r2 = 0.52, F1, 6 = 6.52, P = 0.04, y = 40.03 + 16.94x).   However,  although community dissimilarity showed a slight  increase with  increasing number of  animals  per  treatment,  this  was  not  significant  (F1,  6  =  3.56,  P  =  0.11).    Likewise,  rarefied  richness was  not  related  to  disturbance  (F1, 6  =  0.06,  P  =  0.82).    Resource  heterogeneity  (productivity  variation)  within  treatments  did  not  increase  with  disturbance  (F1,  6  =  0.15,  P  =  0.71).    Taxonomic  richness  increased with  increasing  variation in productivity but was only significant at the 1 % level (r2 = 0.47, F1, 6 = 5.35,  P = 0.06, y = 15.84 + 3.08x).   Taxonomic dissimilarity was not  related  to variation  in  productivity (F1, 6 = 2.04, P = 0.2).    Overall community structure differed between the four disturbance treatments  (R  =  0.06,  P  =  0.03;  Fig  2a).   However  differences  between  disturbance  treatments  were  largely due to shifts  in density of the most abundant taxa rather than change in  composition  (SIMPER;  Fig.  2;  Fig.  3).    Five  taxa  contributed  on  average  77  %  to  differences  between  disturbance  treatments.    These were:  the  snail  Potamopyrgus  antipodarum, Oligochaeta, Platyhelminthes,  the mayfly Deleatidium spp. and  the net  spinning caddisfly Aoteapsyche colonica.   The caddisfly Pycnocentrodes aeris was one  of the most abundant taxa at all sites and treatments (7.3 – 27.6 % contribution) but  contributed little to differences between sites and disturbance treatments.  Of the top  five taxa differentiating treatments, P. antipodarum  (Fig. 2c) and A. colonica  (Fig. 2d)  were the only taxa to respond linearly to disturbance at the between‐stream scale and  were negatively correlated with each other at both the within‐stream scale (r = ‐0.45, P  Chapter 6: Scale dependent links between productivity, disturbance and diversity      = 0.03) an Fig. 3).  Ho disturbanc   Figure 2: N on  raw  inv Zealand str upright  tria days), squa (every 4 da EX2, and sq Aoteapsych   Co different t dist. 4: R  structure  2/dist.  4:  disturbanc d between wever, wh e treatmen on‐metric m ertebrate  c eams, Dece ngles = dist res = distur ys).    (b) Co uares = EX3 e colonica.   mmunity  s o all other  = 0.12, P = between  th R  =  0.05,  e  treatme ‐stream sc en assessin t were site ultidimensi ommunty  d mber 2009  urbance  rat bance rate 3 ded on  indiv .  (c) Bubble Bubble scale tructure  o treatments  0.01; Fig.  e  remainin P  =  0.09;  d nts  indicat 1 ale althoug g abundan  specific an onal scaling  ata  from  a (stress = 0.1 e 1  (not dis  (every 8 da idual stream  plot for ab  for both =  f  undisturb  (dist. 2: R 2b).   Howe g  treatme ist.  3/dist. ed  the  nu   59  h only at t ces at each d varied gr (NMDS) ord rtificial  strea 3).   (a) Code turbed),  cir ys), and inv s: upright t undance of  0 – 300 indiv ed  treatm = 0.11, P =  ver, there  nts  (dist. 2   4:  R  =  ‐0. mber  of  t he 1 %  leve  individual  eatly.    l (r =  ‐0.67 site, the re , P = 0.07; sponses to ination of a m  channels d based on cles = distur erted triang riangles = E Potamopyrg iduals per s ents  (dist.  0.03; dist.  was no diff /dist. 3: R  01,  P  =  0. axa  contri xis one agai   in  Hawke’  disturbance bance  rate  les = disturb X1,  inverted us antipoda ample.    1)  were  s 3: R = 0.08, erence  in c = 0.02, P = 52).    Varia buting  to        nst axis two s  Bay,  New  treatment: 2  (every 16 ance rate 4  triangles = rum and (d)               ignificantly  P = 0.047; ommunity  0.22; dist. tion within differences             Chapter 6: Scale dependent links between productivity, disturbance and diversity      160    between  replicates  declined with  disturbance  rate.    Four  taxa  (P.  antipodarum,  A.  colonica, Pycnocentrodes aeris and Deleatidium spp.) each contributed greater than 10  % to differences at the lowest disturbance rates whereas P. antipodarum was the only  taxa to contribute greater than 10 % at high disturbance sites.  The contribution of P.  antipodarum to differences between replicates increased with disturbance rate.        Figure  3:  Mean  abundance  of  five  taxa  contributing  the  greatest  to  difference  between  disturbance  treatments  at each disturbance  treatment  collected within  artificial  channels  in  three streams (EX1 – EX3) in Hawke’s Bay, New Zealand, December 2009.    Chapter 6: Scale dependent links between productivity, disturbance and diversity      161    P.  antipodarum was  the most  abundant  taxa  throughout  all  treatments  but  abundance  of  this  snail was  lowest  at  the  least  disturbed  treatments  (N  =  44)  and  highest  at  the most  disturbed  (N  =  180)  and  average  per  treatment  increased with  disturbance  rate  (r2 = 0.55, F1, 6 = 7.34, P = 0.04, y = 11.75 + 39.77x; Fig. 2c; Fig. 3).   However,  differences  in  P.  antipodarum  were  not  significant  between  disturbance  treatments (F3, 20 = 0.98, P = 0.42).  A. colonica declined monotonically with increasing  disturbance (r2 = 0.79, F1, 6 = 22.37, P = 0.003, y = 28.42 – 6.3x).   P. antipodarum was  negatively correlated with both axis 1  (r =  ‐0.67) and axis 2  (r =  ‐0.45) of  the NMDS  ordination  (Fig. 2c).   A.  colonica was positively  correlated with  axis 1  (r = 0.58)  and  negatively correlated with axis 2 (r = ‐0.36; Fig. 2d).  P. antipodarum was highest at site  EX2 (mean = 247 ind./sample) and lowest at site EX1 (mean = 11 ind./sample) (F2, 21 =  12.6, P = 0.0003).  A. colonica was highest at EX1 (mean = 34 ind./sample) and lowest  at both EX2 and EX3 (mean = 2 ind./sample) (F2, 21 = 15.57, P = 0.0001).    The difference in community structure was more strongly related to individual  site differences in composition (R = 0.65, P = 0.001; Fig. 2b).  Communities were more  similar at each site regardless of disturbance rate (Bray‐Curtis similarity = 46.6 – 50.1)  than between disturbance treatments at all sites (Bray‐Curtis similarity = 28.6 – 40.7).   Moreover,  when  assessing  differences  in  community  composition  using  presence/absence  data,  communities  were more  similar  at  each  site  regardless  of  disturbance  rate  (Bray‐Curtis  similarity  =  61.5  –  68.3)  than  between  disturbance  y‐Curtis similarity = 54.7 ‐ 57).   treatments at all sites (Bra   Discussion  The response of diversity to both productivity and disturbance varied with the spatial  scale at which  it was examined  in  this experiment.   This  is a  common phenomenon  when  assessing  these  relationships  (Petraitis  et  al.  1989,  Currie  1991, Mackey  and  Currie  2001,  Chase  and  Leibold  2002).    In  fact,  diversity  only  responded  to  one  of  productivity or disturbance at each scale.  Productivity affected taxonomic richness at  within‐  (local)  and  between‐stream  (regional)  scales  and  disturbance  affected  community dissimilarity  (β diversity).   Previous  studies have  shown  that productivity  Chapter 6: Scale dependent links between productivity, disturbance and diversity      162    and disturbance can both affect diversity (e.g. Tonkin and Death In prep., Tonkin et al.  In prep.).   Although central models  in community ecology  such as  the  IDH and DEM  predict the disturbance‐diversity relationship  is humped (Grime 1973b, Connell 1978,  Huston  1979,  1994,  Kondoh  2001),  there  is  a  large  body  of  evidence  to  suggest  diversity  simply  declines with  increasing  disturbance  (Death  and Winterbourn  1995,  Vinson and Hawkins 1998, Mackey and Currie 2001, Death 2002), but this relationship  can vary with spatial and temporal scale (Petraitis et al. 1989).  We predicted diversity  in  these  streams  at  all  scales would decline with  increasing  rate of  disturbance but  there was no effect of disturbance at within‐ and between‐stream  scales.   Although  experimental  studies  have  struggled  to  isolate  the  effects  of  disturbance  and  productivity  (Robinson and Minshall 1986, Death 1996), we were able  to manipulate  productivity  independent  of  disturbance  as  chlorophyll  a  remained  higher  in  high  productivity  treatments.    The  fact  that productivity  and disturbance  are particularly  linked  in  lotic systems suggests conclusions drawn  from this paper are not  limited to  these systems (Lake 2000, Death 2002).      Productivity­diversity relationship  The  likelihood  of  occurrence  of  u‐shaped  relationships  between  productivity  and  diversity,  as  found  here  at  the  within‐stream  scale,  increases  with  spatial  scale  (Mittelbach  et  al.  2001).    Identifying  causes  for  this  pattern  is  difficult  as  it  is  not  supported by theory and the mechanisms have not been discussed (Mittelbach et al.  2001).   Most emphasis  is placed on  identifying humped trends as much of ecological  theory  predicts  this  (Mittelbach  et  al.  2001),  but  there  are  reasons  to  expect  other  forms of this relationship (Abrams 1995).  The more common unimodal relationship at  local  scales  (e.g. Grime  1973a, Huston  1979,  Rosenzweig  1992,  1995,  Leibold  1999,  Waide  et  al.  1999, Mittelbach  et  al.  2001)  is  likely  to  involve  competition  between  animals  within  heterogeneous  habitats  (Kassen  et  al.  2000),  but  not  relatively  homogeneous such as  in our study.   This requires animals’ competitive ability to vary  between  niches which  in  turn would  require  a  trade‐off  between  competitive  and  colonising ability which is rare in streams (Death 2002).    Chapter 6: Scale dependent links between productivity, disturbance and diversity      163    As  predicted,  between‐stream  richness  increased  monotonically  with  increasing  productivity.    The  discrepancy  in  the  productivity‐diversity  relationship  found when assessing within‐ and between‐stream  scales  in  this  study matches  that  commonly  found  between  true  local  and  regional  scales  (Currie  1991,  Chase  and  Leibold 2002).  Chase and Leibold (2002) found diversity peaked at intermediate levels  of  productivity  at  small  scales  and  increased monotonically  at  the  regional  scale  in  pond communities.   Several mechanisms have been put forward to explain the  linear  increase in richness at regional scales (Currie 1991, Waide et al. 1999, Mittelbach et al.  2001) and one likely factor is the inclusion of different niches as spatial scale increases  (Kassen et al. 2000).   Along with  the  regional  linear  increase  in  richness, Chase and  Leibold (2002) found a monotonic  increase  in community dissimilarity with  increasing  productivity and more recently Chase (2010) suggested a greater deterministic control  in low productivity environments.  Many factors could explain this including increased  resource  heterogeneity  and  a  greater  number  of  stable  states  (Chase  and  Leibold  2002), but due to the nature of this experiment are not likely to apply.  Accordingly, we  found no relationship between community dissimilarity and productivity.      Dissimilarity increased with disturbance   Increasing disturbance led to greater dissimilarity between communities; this counters  our  prediction  of  greater  deterministic  control  at  higher  disturbance  (Chase  2003,  2007).  Often differences in community dissimilarity can be explained by differences in  the recolonisation process; i.e. deterministic and stochastic assembly patterns.  Theory  predicts β diversity should decline with environmental harshness due to a greater role  of deterministic assembly (Chase 2007), such as that found when comparing spring and  runoff‐fed stream communities (Barquin and Death 2006).  Nevertheless, recent work  in  lotic  systems  has  demonstrated  a  greater  deterministic  control  at  intermediate  levels  of  disturbance  (Lepori  and  Malmqvist  2009)  suggesting  the  transition  from  stochastic to deterministic control  is not necessarily  linear.   Disturbance, especially  in  streams,  typically operates at a patchy  scale and creates heterogeneous habitat and  resources  (Doeg  et  al.  1989,  Lancaster  and  Hildrew  1993).    Thus,  disturbance  can  promote diversity within a  landscape  (β and  γ) by allowing both early colonising and  Chapter 6: Scale dependent links between productivity, disturbance and diversity      164    late  successional  species  to  coexist, but  this  is dependent on  low dispersal  (Warren  1996,  Chase  2003,  Chase  and  Ryberg  2004, Ostman  et  al.  2006, Matthiessen  et  al.  2010).    Consequently,  in  natural  systems  we  could  have  expected  the  increased  dissimilarity  with  disturbance  to  be  a  result  of  increased  resource/habitat  heterogeneity.  Due to the spatial and temporal scale of this experiment (dissimilarity  assessed  between  equal  disturbance  treatments),  disturbance  was  homogeneous  between streams.   Accordingly, resource heterogeneity (productivity variation) did not  increase  with  disturbance  rate  nor  did  community  dissimilarity  with  resource  heterogeneity.    In  fact,  resource  heterogeneity  is  likely  to  be  higher  in more  stable  habitats at this scale (Beisel et al. 1998, Barquin and Death 2006).    Whether  the  increase  in  dissimilarity with  disturbance  is  a  result  of  greater  stochastic assembly  in more disturbed  treatments  is unclear  (Chase 2003, 2007).   A  large proportion of New Zealand benthic  invertebrates are capable of  living  in highly  disturbed environments as a result of the high degree of environmental stochasticity  and have generalist feeding habits as a result (Winterbourn et al. 1981, Thompson and  Townsend  2000).    In  fact,  there  are many  that  could  be  considered  ‘Hutchinsonian  demons’ – both superior colonisers and competitors (Kneitel and Chase 2004, Cadotte  et al. 2006).   Essentially the role of deterministic assembly  is  less pronounced than  in  other  ecosystems where  harsh  conditions  select  for  few  specialised  taxa.    In  these  streams,  the  proportion  of  taxa  within  the  regional  pool  able  to  withstand  these  conditions is likely to be high.    Given  the hypothesised  lack of deterministic control  in disturbed  treatments,  we  suggest  the  increase  in community dissimilarity with disturbance  rate  is  simply a  function  of  the  time  since  last  disturbance  (Reice  1994).    Considered  this  way,  community  dissimilarity  declined with  time,  thus  it  appears  these  communities  are  converging  on  a  single  stable  equilibrium.    If  β  diversity  is  not  promoted  through  different  assembly  processes  (multiple  stable  equilibria),  then  heterogeneity  in  resources/conditions  between  streams  (with  single  stable  equilibria)  would  be  required  (Loreau and Mouquet 1999, Mouquet and  Loreau 2002).    Low  connectivity  between  replicates  as  in  this  study  (dispersal  between  streams),  typically  promotes  multiple  states  (Warren  1996,  Chase  2003,  Chase  and  Ryberg  2004,  Ostman  et  al.  Chapter 6: Scale dependent links between productivity, disturbance and diversity      165    2006,  Matthiessen  et  al.  2010).    Therefore,  the  convergence  on  a  single  stable  equilibrium  with  time  since  disturbance  between  the  streams  indicates  little  heterogeneity  between  habitats  and  resources  which  was  indeed  evident  in  the  present study.      Local suite of colonists  Rather than differences in assembly rules, it appears colonisation patterns are simply a  function  of  the  local  suite  of  colonising  taxa.    Specifically,  colonisation  of  channels  appears  to  have  come  from  the  small‐scale  dispersal  of  animals  present within  the  surrounding  benthos  of  the  spring‐fed  streams  in which  the  study was  undertaken;  rather than a specialist suite of colonist taxa.  Previous experiments of this nature have  found this pattern of local arrival rather than the more broad‐scale arrival of specialist  colonisers (Death 2006).   Thus the findings need to be considered with some caution  as  the  colonisation  patterns  do  not  necessarily match  those  of  large‐scale  natural  disturbances.    Communities  found within  the  experimental  channels, with  a  strong  contribution  of  Potamopyrgus  antipodarum,  Aoteapsyche  colonica,  Platyhelminthes,  Oligochaeta, Deleatidium spp. and Pycnocentrodes aeris are similar to those  found  in  the benthos in a previous study on these streams around this time (Tonkin and Death  In prep.).   If ecological theory holds, we would expect high disturbance treatments to  be  the most  similar  due  to  deterministic  control  selecting  for  few  capable  taxa  but  communities were more similar within streams than within disturbance treatments.    Our results suggest the increasing dissimilarity with disturbance rate was due to  shifts  in abundance of the most dominant taxa, although the mechanisms underlying  this  are  unclear.    The most  abundant  animal,  the  hydrobiid  snail  P.  antipodarum,  increased with disturbance  rate which opposes  research  suggesting  that  this  snail  is  inversely  related  to disturbance  in  streams  (Holomuzki  and Biggs 1999).   Moreover,  molluscs  are  generally  considered  slower  colonisers  and  likely  to  be  competitive  dominants  (Hemphill  and Cooper  1983, Mackay  1992), but  the  atypical  response  to  disturbance  in  this  instance  almost  certainly  reflects  their  abundance  in  the  surrounding benthos.  The response of P. antipodarum to disturbance has been found  to  be  dependent  on  the  type  of  substrate  in  the  streambed  (Holomuzki  and  Biggs  Chapter 6: Scale dependent links between productivity, disturbance and diversity      166    1999), but this  is not  likely a factor  in the present study as substrates were relatively  uniform  between  treatments.    The  decline  of  P.  antipodarum with  time  since  last  disturbance may have been due  to  a  competitive  interaction with  the hydropsychid  caddisfly A.  colonica.   Disturbance  can  regulate  the  rate of  competitive  interactions  between  opportunistic  early  colonisers  and  competitive  dominants  which  colonise  slower (Hemphill and Cooper 1983).  However, A. colonica did not necessarily replace  P.  antipodarum  in  this  study,  merely  their  relative  abundances  increased  and  decreased respectively.  Although the negative interaction between these two species  occurred within each stream, densities were clearly more site specific than disturbance  specific.  Nonetheless, where the substrate is regularly being turned over, it is likely to  be  unfavourable  conditions  for  a  net  building  caddisfly  such  as  A.  colonica  which  supports their increase with time since disturbance.   No suite of taxa were replaced with time since disturbance (Collier and Quinn  2003,  Death  2006),  but  the  negative  association  between  A.  colonica  and  P.  antipodarum  indicates  a  potential  competitive‐colonising  trade‐off.    A  critical  requirement of many  relationships  in disturbance ecology  is  that  there  is a  trade‐off  between competitive and colonising ability (Chesson and Huntly 1997, Roxburgh et al.  2004, Cadotte 2007).   Neutral models have  renewed debate on  this central  tenet  in  ecology (e.g. Hubbell 2001), but there  is still widespread support for these trade‐offs  (e.g.  Kneitel  and  Chase  2004,  Cadotte  et  al.  2006).    Previous  studies  looking  at  responses disturbance  in streams  indicate that rather than competitive displacement,  original  taxa  remain  and  other  taxa  arrive  with  time  but  do  not  displace  early  colonisers  (Death 1996, Collier and Quinn 2003, Death 2006).   The expectation of no  competitive displacement  led us to predict a decline  in richness with disturbance but  we  found  no  such  pattern.    Chase  (2010)  also  found  that  taxa  in  low  productivity  habitats  (harsh) were  simply a nested  subset of  those  found at higher productivities  (benign), and argue that this is likely a more general pattern than previously thought.   Indeed, the taxa found in high disturbance treatments in the present study are merely  a  subset of  those  found  in  stable  treatments  and  although P. antipodarum declines  with  increasing  stability,  it  is  still  dominant  throughout  all  treatments.    More  Chapter 6: Scale dependent links between productivity, disturbance and diversity      167    importantly,  an  underlying  competitive  interaction  between  P.  antipodarum  and  A.  colonica does not explain the increase in dissimilarity with disturbance.      Conclusions  Our  results  demonstrate  that  spatial  scale  can  affect  the  way  species  diversity  responds to environmental conditions.   We have shown that diversity can respond to  both  productivity  and  disturbance  but  at  different  spatial  scales.    Specifically,  taxonomic  richness was  lowest  at  intermediate  levels  of  productivity  at  the within‐ stream  (local)  scale  and  increased monotonically  at  the  between‐stream  (regional)  scale.    Rather  than  the  increase  in  community  dissimilarity  previously  found  with  increasing productivity (Chase and Leibold 2002) and the stronger role of deterministic  assembly  suggested  for  harsh  environments  (Chase  2007), we  found  an  increase  in  dissimilarity with  increasing  rate of disturbance.   Due  to  the  abilities of  the pool of  colonists  to  withstand  disturbed  environments  in  these  streams,  communities  are  converging as a  function of the time since  last disturbance.   Namely, they are simply  returning  to  those  found  in  the  surrounding  benthos  as  a  result  of  the  scale  of  disturbance  and  that  it  is  not  limiting  the  species  pool.    These  patterns  are  highly  dependent on the pool of colonists and are underpinned by  interactions within each  stream.    Without  a  distinction  between  early  colonising  and  late  successional  communities,  increased  deterministic  assembly  (and  lowering  of  β  diversity)  with  increasing  disturbance  will  not  apply.    In  fact,  depending  on  the  heterogeneity  of  regional habitats and connectivity between habitats, the opposite pattern may occur  as we show here.      References  Abrams,  P. A.  1995. Monotonic  or  unimodal  diversity  productivity  gradients  ‐ What  does competition theory predict. Ecology 76:2019‐2027.  Barquin,  J., and R. G. Death. 2006. Spatial patterns of macroinvertebrate diversity  in  New Zealand springbrooks and rhithral streams. Journal of the North American  Benthological Society 25:768‐786.  Chapter 6: Scale dependent links between productivity, disturbance and diversity      168    Beisel,  J.  N.,  P.  Usseglio‐Polatera,  S.  Thomas,  and  J.  C.  Moreteau.  1998.  Stream  community  structure  in  relation  to  spatial  variation:  the  influence  of  mesohabitat characteristics. Hydrobiologia 389:73‐88.  Cadotte, M. W. 2007. Competition‐colonization  trade‐offs and disturbance effects at  multiple scales. Ecology 88:823‐829.  Cadotte, M. W., D. V. Mai, S. Jantz, M. D. Collins, M. Keele, and J. A. Drake. 2006. On  testing  the  competitioncolonization  tradeoff  in  a  multispecies  assemblage.  American Naturalist 168:704‐709.  Cardinale, B. J., H. Hillebrand, and D. F. Charles. 2006. Geographic patterns of diversity  in  streams  are  predicted  by  a  multivariate  model  of  disturbance  and  productivity. Journal of Ecology 94:609‐618.  Chase,  J.  M.  2003.  Community  assembly:  when  should  history  matter?  Oecologia  136:489‐498.  Chase,  J.  M.  2007.  Drought  mediates  the  importance  of  stochastic  community  assembly. Proceedings of the National Academy of Sciences 104:17430‐17434.  Chase, J. M. 2010. Stochastic Community Assembly Causes Higher Biodiversity in More  Productive Environments. Science 328:1388‐1391.  Chase,  J.  M.,  and  M.  A.  Leibold.  2002.  Spatial  scale  dictates  the  productivity‐ biodiversity relationship. Nature 416:427‐430.  Chase,  J.  M.,  and  W.  A.  Ryberg.  2004.  Connectivity,  scale‐dependence,  and  the  productivity‐diversity relationship. Ecology Letters 7:676‐683.  Chesson, P., and N. Huntly. 1997. The roles of harsh and fluctuating conditions  in the  dynamics of ecological communities. The American Naturalist 150:519‐553.  Clarke,  K.  R.,  and  R. N. Gorley.  2006.  PRIMER  v6: User Manual/Tutorial.  PRIMER‐E,  Plymouth.  Collier, K. J., and J. M. Quinn. 2003. Land‐use influences macroinvertebrate community  response following a pulse disturbance. Freshwater Biology 48:1462‐1481.  Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs. Science 199:1302‐ 1310.  Currie, D. J. 1991. Energy and large‐scale patterns of animal‐species and plant‐species  richness. American Naturalist 137:27‐49.  Chapter 6: Scale dependent links between productivity, disturbance and diversity      169    Currie,  D.  J.,  and  V.  Paquin.  1987.  Large‐scale  biogeographical  patterns  of  species  richness of trees. Nature 329:326‐327.  Death, R. G. 1996. The effect of patch disturbance on stream invertebrate community  structure: the influence of disturbance history. Oecologia 108:567‐576.  Death, R. G. 2002. Predicting invertebrate diversity from disturbance regimes in forest  streams. Oikos 97:18‐30.  Death,  R.  G.  2006.  Colonisation  in  New  Zealand  streams:  predictable  patterns  or  chance events? New Zealand Natural Sciences 31:93‐112.  Death,  R.  G.,  and  M.  J.  Winterbourn.  1995.  Diversity  patterns  in  stream  benthic  invertebrate communities:  the  influence of habitat  stability. Ecology 76:1446‐ 1460.  Doeg, T. J., P. S. Lake, and R. Marchant. 1989. Colonization of experimentally disturbed  patches  by  stream  macroinvertebrates  in  the  Acheron  River,  Victoria.  Australian Journal of Ecology 14:207‐220.  Fukami, T., and P. J. Morin. 2003. Productivity‐biodiversity relationships depend on the  history of community assembly. Nature 424:423‐426.  Gaston, K. J. 2000. Global patterns in biodiversity. Nature 405:220‐227.  Graham, A. A., D. J. McCaughan, and F. S. McKee. 1988. Measurement of surface area  of stones. Hydrobiologia 157:85‐87.  Grime,  J. P. 1973a. Competitive exclusion  in herbaceous vegetation. Nature 242:344‐ 347.  Grime,  J.  P.  1973b.  Control  of  species  density  in  herbaceous  vegation.  Journal  of  Environmental Management 1:151‐167.  Hart,  D.  D.  1992.  Community  organization  in  streams:  the  importance  of  species  interactions, physical factors, and chance. Oecologia 91:220‐228.  Hastings,  A.  1980.  Disturbance,  coexistence,  history,  and  competition  for  space.  Theoretical Population Biology 18:363‐373.  Hemphill, N., and S. D. Cooper. 1983. The effect of physical disturbance on the relative  abundances of 2 filter‐feeding insects in a small stream. Oecologia 58:378‐382.  Holomuzki, J. R., and B. J. F. Biggs. 1999. Distributional responses to flow disturbance  by a stream‐dwelling snail. Oikos 87:36‐47.  Chapter 6: Scale dependent links between productivity, disturbance and diversity      170    Hubbell,  S.  P.  2001.  The  unified  neutral  theory  of  biodiversity  and  biogeography.  Princeton University Press.  Huston, M. 1979. A general hypothesis of  species diversity. The American Naturalist  113:81‐100.  Huston,  M.  1994.  Biological  Diversity:  The  Coexistence  of  Species  on  Changing  Landscapes. Cambridge University Press, Cambridge, UK.  Kassen, R., A. Buckling, G. Bell, and P. B. Rainey. 2000. Diversity peaks at intermediate  productivity in a laboratory microcosm. Nature 406:508‐512.  Kneitel, J. M., and J. M. Chase. 2004. Trade‐offs  in community ecology:  linking spatial  scales and species coexistence. Ecology Letters 7:69‐80.  Kondoh, M.  2001. Unifying  the  relationships  of  species  richness  to  productivity  and  disturbance.  Proceedings  of  the  Royal  Society  of  London  Series  B‐Biological  Sciences 268:269‐271.  Lake, P. S. 2000. Disturbance, patchiness, and diversity in streams. Journal of the North  American Benthological Society 19:573‐592.  Lancaster, J., and A. G. Hildrew. 1993. Flow Refugia and the Microdistribution of Lotic  Macroinvertebrates.  Journal  of  the  North  American  Benthological  Society  12:385‐393.  Leibold,  M.  A.  1999.  Biodiversity  and  nutrient  enrichment  in  pond  plankton  communities. Evolutionary Ecology Research 1:73‐95.  Lepori,  F.,  and  B. Malmqvist.  2009.  Deterministic  control  on  community  assembly  peaks at intermediate levels of disturbance. Oikos 118:471‐479.  Loreau, M., and N. Mouquet. 1999. Immigration and the maintenance of local species  diversity. American Naturalist 154:427‐440.  Mackay, R. J. 1992. Colonization by lotic macroinvertebrates: a review of processes and  patterns. Canadian Journal of Fisheries and Aquatic Science 49:617‐628.  Mackey,  R.  L.,  and  D.  J.  Currie.  2001.  The  diversity‐disturbance  relationship:  Is  it  generally strong and peaked? Ecology 82:3479‐3492.  Matthiessen,  B.,  E. Mielke,  and  U.  Sommer.  2010.  Dispersal  decreases  diversity  in  heterogeneous metacommunities by enhancing  regional competition. Ecology  91:2022‐2033.  Chapter 6: Scale dependent links between productivity, disturbance and diversity      171    Mittelbach, G. G., C. F. Steiner, S. M. Scheiner, K. L. Gross, H. L. Reynolds, R. B. Waide,  M. R. Willig, S. I. Dodson, and L. Gough. 2001. What is the observed relationship  between species richness and productivity? Ecology 82:2381‐2396.  Morin, A., W. Lamourex, and  J. Busnarda. 1999. Empirical models predicting primary  productivity  from chlorophyll a and water temperature  for stream periphyton  and  lake  and  ocean  phytoplankton.  Journal  of  the  North  American  Benthological Society 18:299‐307.  Mouquet, N.,  and M.  Loreau.  2002.  Coexistence  in metacommunities:  The  regional  similarity hypothesis. American Naturalist 159:420‐426.  Ostman, O., J. M. Kneitel, and J. M. Chase. 2006. Disturbance alters habitat isolation's  effect on biodiversity in aquatic microcosms. Oikos 114:360‐366.  Petraitis, P. S., R. E. Latham, and R. A. Niesenbaum. 1989. The Maintenance of Species‐ Diversity by Disturbance. Quarterly Review of Biology 64:393‐418.  Pickett, S. T. A., and P. S. White. 1985. The ecology of natural disturbance and patch  dynamics. Academic Press.  Poff,  N.  L.  1997.  Landscape  filters  and  species  traits:  Towards  mechanistic  understanding and prediction in stream ecology. Journal of the North American  Benthological Society 16:391‐409.  Reice,  S.  R.  1994.  Nonequilibrium  determinants  of  biological  community  structure.  American Scientist 82:424‐435.  Robinson, C. T., and G. W. Minshall. 1986. Effects of disturbance frequency on stream  benthic community structure in relation to canopy cover and season. Journal of  the North American Benthological Society 5:237‐248.  Rosenzweig, M. L. 1992. Species‐diversity gradients ‐ we know more and less than we  thought. Journal of Mammalogy 73:715‐730.  Rosenzweig, M.  L.  1995.  Species Diversity  in  Space  and  Time.  Cambridge University  Press, Cambridge, UK.  Rosenzweig, M. L., and Z. Abramsky. 1993. How are Diversity and Productivity Related?  Pages  52‐65  in  R.  E.  Ricklefs  and  D.  Schluter,  editors.  Species  Diversity  in  Biological Communities. University of Chicago Press, Chicago, Illinois, USA.  Chapter 6: Scale dependent links between productivity, disturbance and diversity      172    Roxburgh,  S.  H.,  K.  Shea,  and  J.  B.  Wilson.  2004.  The  intermediate  disturbance  hypothesis:  Patch  dynamics  and mechanisms  of  species  coexistence.  Ecology  85:359‐371.  Sousa, W. P. 1979. Disturbance in marine intertidal boulder fields: the nonequilibrium  maintenance of species diversity. Ecology 60:1225‐1239.  Steinman, A. D.,  and G. A.  Lamberti.  1996. Biomass  and  pigments  of  benthic  algae.  Pages 295‐314  in F. R. Hauer and G. A. Lamberti, editors. Methods  in Stream  Ecology. Academic Press, San Diego, CA.  Thompson, R., and C. Townsend. 2006. A truce with neutral theory: local deterministic  factors,  species  traits and dispersal  limitation  together determine patterns of  diversity in stream invertebrates. Journal of Animal Ecology 75:476‐484.  Thompson,  R. M.,  and  C.  R.  Townsend.  2000.  New  Zealand's  stream  invertebrate  communities:  an  international  perspective.in  K.  J.  Collier  and  M.  J.  Winterbourn,  editors.  New  Zealand  stream  invertebrates:  ecology  and  implications for management. New Zealand Limnological Society, Hamilton.  Tonkin,  J.  D.,  and  R.  G.  Death.  In  prep.  Consistent  effects  of  productivity  and  disturbance on diversity between landscapes.  Tonkin,  J. D., R. G. Death, and K.  J. Collier.  In prep. Do productivity and disturbance  interact to modulate macroinvertebrate diversity in streams?  Towns,  D.  R.,  and  W.  L.  Peters.  1996.  Leptophlebiidae  (Insecta:  Ephemeroptera).  Manaaki Whenua Press, Lincoln, New Zealand.  Vinson, M. R., and C. P. Hawkins. 1998. Biodiversity of stream insects: variation at local,  basin, and regional scales. Annual Review of Entomology 43:271‐293.  Waide, R. B., M. R. Willig, C.  F.  Steiner, G. Mittelbach,  L. Gough,  S.  I. Dodson, G. P.  Juday,  and  R.  Parmenter.  1999.  The  relationship  between  productivity  and  species richness. Annual Review of Ecology and Systematics 30:257‐300.  Warren,  P.  H.  1996.  Dispersal  and  destruction  in  a  multiple  habitat  system:  An  experimental approach using protist communities. Oikos 77:317‐325.  Winterbourn, M.  J., K.  L. D. Gregson, and C. H. Dolphin. 2000. Guide  to  the aquatic  insects of New Zealand. Entomological Society of New Zealand, Auckland.  Chapter 6: Scale dependent links between productivity, disturbance and diversity      173    Winterbourn,  M.  J.,  J.  S.  Rounick,  and  B.  Cowie.  1981.  Are  New  Zealand  stream  ecosystems  really  different? New  Zealand  Journal  of Marine  and  Freshwater  Research 15:321‐328.                    Chapter 7  Synthesis                        Chapter 7: Synthesis      175    The  aim  of  this  thesis  was  to  investigate  the  linkages  between  productivity,  disturbance  and diversity  in  stream macroinvertebrate  communities  and  further  the  understanding  of  these  relationships.    I  accomplished  this  by  (i)  assessing  the  application of diversity models such as the Dynamic Equilibrium Model (DEM) (Huston  1979,  1994,  Kondoh  2001)  and  a more  contemporary  stream‐specific  productivity‐ disturbance‐diversity  model  (Death  2002),  as  well  as  developing  and  testing  a  multivariate model of diversity based on observed patterns in New Zealand mountain  streams; and (ii) exploring the form these links in further detail between New Zealand  and Spanish streams as well as investigating the underlying mechanisms driving these  relationships  in  lotic systems.   The prevailing premise  is  that diversity  increases with  increasing productivity and declines with  increasing disturbance, although  there was  slight variation in the form of these relationships between chapters.    The  strength  of  the  link  between  productivity,  disturbance  and  diversity  of  stream  macroinvertebrate  communities  was  clearly  demonstrated  in  this  study.   Firstly,  a  large  scale  study  of  mountain  streams  indicated  that  productivity  and  disturbance  could  both  independently  predict  diversity  but  the  relationship  was  affected by  the presence of canopy cover,  in  that  strong  links were present  in open  canopy  streams  but  not  those  with  canopy  cover  (Chapter  2).    However,  this  relationship  was  consistent  between  regions  of  differing  land  use  (Chapter  3).   Secondly,  I  demonstrate  that  where  productivity‐diversity  relationships  are  not  obvious,  invertebrate  communities  may  be  controlled  by  the  growth  form  of  the  periphyton  community  rather  than  its  biomass  (Chapter  4),  or  may  simply  be  a  function  of  the  range  of  productivity  considered  (Chapter  5).    In  fact,  global  productivity‐diversity  relationships  may  exist  in  streams,  even  where  individual  differences  in the shape of the curve between regions are evident.   Finally, there  is a  strong  effect  of  the  spatial  scale  of  consideration  in  productivity‐diversity  and  disturbance‐diversity relationships (Chapter 6).     One of the more  important hypotheses I tested was whether productivity and  disturbance interact to modulate diversity (Kneitel and Chase 2004, Scholes et al. 2005,  Cardinale  et  al.  2006),  or  whether  they  operate  independently.    Because  the  link  between  productivity  and  disturbance  in  streams  is  particularly  strong  (Lake  2000,  Chapter 7: Synthesis      176    Death 2002, Cardinale et al. 2006), they are thought to interact to generate outcomes  on diversity.   This  interactive effect  is a central component of the DEM, but although  both productivity and disturbance could predict diversity at individual scales, there was  no evidence of a statistical interaction between the two (Chapter 2 and 3).  That is, the  effects  of  productivity  and  disturbance were  additive,  rather  than multiplicative  at  individual scales.    In Chapter 2 I found some support for the Intermediate Disturbance Hypothesis  (IDH) (Grime 1973, Connell 1978, Sousa 1979), with a unimodal relationship between  disturbance  and  diversity.    However,  a  second  study  assessing  this  relationship  suggested  diversity  declined with  disturbance  and  this was  consistent  between  two  regions of contrasting  land use  intensity (Chapter 3).   Despite finding support for the  IDH  in Chapter 2,  the  relationship between productivity and diversity did not match  that  of  the  DEM,  exhibiting  a  log‐linear  increase  in  diversity  with  increasing  productivity.    This  increase matches  the  assertion  of Death  (2002)  that productivity  sets the upper limit to richness in stream communities rather than leading to a greater  rate of competitive interactions.  Death (2002) further suggests the role of disturbance  is  to  simply  remove  taxa  and  reset  colonisation,  allowing  for  productivity  to  set  maximum richness.  Therefore, I developed a model, building on Death’s model, which  predicts  that  diversity  is  a  unimodal/quadratic  function  of  disturbance,  with  productivity determining the magnitude of that peak.      This model, and that of Death (2002), were similar  in their predictive ability  in  both a pristine region and a region more heavily impacted by land use change (Chapter  3).  However, this suggested a quadratic decline rather than the unimodal trend found  in Chapter 2.   The  link between productivity, disturbance and diversity did not differ  between these regions, which is interesting considering the contrast in anthropogenic  influences  in each  region’s environment.   When assessing whether  the productivity‐ disturbance‐diversity  relationship  differed  between  streams with  or without  canopy  cover,  I  found  it was only evident at open canopy sites.   This was not a result of the  correlation between canopy cover and stream size and  it remains unclear as to what  the  main  determinant  of  diversity  in  sites  with  canopy  cover  is.    However,  the  difference  in productivity‐disturbance‐diversity  links between  sites with and without  Chapter 7: Synthesis      177    canopy is likely to be related to a stronger coupling between invertebrate communities  and  algal productivity  in  autotrophic  than heterotrophic  streams.    The  relationships  between productivity, disturbance and diversity can be highly variable with the spatial  scale at which they are assessed (Petraitis et al. 1989, Currie 1991, Mackey and Currie  2001,  Chase  and  Leibold  2002)  as  I  have  shown  in  experimental  stream  channels  (Chapter 6), and altering productivity or disturbance can  lead to changes  in assembly  sequences of animals.  Interestingly, this experiment indicated that either productivity  or  disturbance  individually  affected  diversity  at  each  scale.    This  differs  from  the  findings  in  Chapters  2  and  3 where,  although  operating  independently,  productivity  and  disturbance  both  had  strong  influences  on  diversity  at  the  same  scale  of  assessment.    Given that productivity‐diversity relationships are a relatively little studied area  of lotic ecology, I was interested in whether productivity set the upper limit to richness  in  streams  elsewhere  in  the world  (Death  2002,  Barquin  2004),  or whether  greater  productivity  led  to  competitive  exclusion.    Although  the  effects  of  productivity  on  diversity were strong in all the New Zealand streams, there was no direct relationship  in the 24 Spanish streams assessed (Chapter 4 and 5).    While  measuring  primary  productivity  indicates  the  rate  of  production  of  energy,  it  does  not  account  for  the  specific  form  of  plants  driving  this  production.   Assessing  the  response  of  macroinvertebrate  diversity  to  two  coarse  measures  of  periphyton community composition  found strong responses where a simple measure  of  primary  productivity  (chlorophyll  a)  could  not  (Chapter  4).    Specifically,  macroinvertebrate diversity tended to be negatively associated with filamentous algal  forms  and  positively  associated with  periphytic mats.   Moreover, when  testing  the  response of these streams along the same axis as 24 New Zealand streams (Chapter 5),  they  were  simply  a  continuation  of  a  log‐linear  curve  found  in  the  New  Zealand  streams,  suggesting  this  may  be  a  universal  pattern.    Despite  the  wide  range  of  productivities  assessed,  there was  no  evidence  for  a  decline  in  diversity  at  greater  rates  of  productivity  which  suggests  competitive  exclusion  is  not  occurring  as  is  required for the DEM to apply.    Chapter 7: Synthesis      178    The  overwhelming  evidence  found  in  this  thesis  is  that  productivity  simply  increases  diversity,  with  no  evidence  of  an  increase  in  competitive  interactions  at  greater  levels of productivity.   The  log‐linear  link between productivity and diversity  was not  limited  to New Zealand streams with a universal pattern emerging between  the Spanish and New Zealand streams.   Furthermore,  I provide  reasons  for a  lack of  obvious link between productivity and diversity including underlying biotic interactions  and the scale at which the relationship  is assessed.   Disturbance, on the other hand,  exhibited more  variable  effects  on  diversity,  although  the  prevailing  theme  is  that  disturbance  simply  removes  taxa  at  a  variable  rate.    This  thesis  goes  a  long  way  towards  identifying  the  true  form of both  the productivity‐diversity and disturbance‐ diversity relationships in streams.  I provide further evidence to suggest the DEM is not  applicable  in  stream  systems, which  is  likely due  to  a  lack of  competitive exclusion.   Finally, although the effects of productivity and disturbance are strong, I propose that  they do not  interact to affect diversity  in stream macroinvertebrate communities but  operate independently.        Chapter 7: Synthesis      References  179    Barquin, J. 2004. Spatial patterns of invertebrate communities in spring and runoff‐fed  streams.  PhD  thesis.  PhD  thesis, Massey  University,  Palmerston  North,  New  Zealand.  Cardinale, B. J., H. Hillebrand, and D. F. Charles. 2006. Geographic patterns of diversity  in  streams  are  predicted  by  a  multivariate  model  of  disturbance  and  productivity. Journal of Ecology 94:609‐618.  Chase,  J.  M.,  and  M.  A.  Leibold.  2002.  Spatial  scale  dictates  the  productivity‐ biodiversity relationship. Nature 416:427‐430.  Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs. Science 199:1302‐ 1310.  Currie, D. J. 1991. Energy and large‐scale patterns of animal‐species and plant‐species  richness. American Naturalist 137:27‐49.  Death, R. G. 2002. Predicting invertebrate diversity from disturbance regimes in forest  streams. Oikos 97:18‐30.  Grime,  J.  P.  1973.  Control  of  species  density  in  herbaceous  vegation.  Journal  of  Environmental Management 1:151‐167.  Huston, M. 1979. A general hypothesis of  species diversity. The American Naturalist  113:81‐100.  Huston,  M.  1994.  Biological  Diversity:  The  Coexistence  of  Species  on  Changing  Landscapes. Cambridge University Press, Cambridge, UK.  Kneitel, J. M., and J. M. Chase. 2004. Disturbance, predator, and resource interactions  alter container community composition. Ecology 85:2088‐2093.  Kondoh, M.  2001. Unifying  the  relationships  of  species  richness  to  productivity  and  disturbance.  Proceedings  of  the  Royal  Society  of  London  Series  B‐Biological  Sciences 268:269‐271.  Lake, P. S. 2000. Disturbance, patchiness, and diversity in streams. Journal of the North  American Benthological Society 19:573‐592.  Mackey,  R.  L.,  and  D.  J.  Currie.  2001.  The  diversity‐disturbance  relationship:  Is  it  generally strong and peaked? Ecology 82:3479‐3492.  Chapter 7: Synthesis      180    Petraitis, P. S., R. E. Latham, and R. A. Niesenbaum. 1989. The Maintenance of Species‐ Diversity by Disturbance. Quarterly Review of Biology 64:393‐418.  Scholes, L., P. H. Warren, and A. P. Beckerman. 2005. The combined effects of energy  and  disturbance  on  species  richness  in  protist  microcosms.  Ecology  Letters  8:730‐738.  Sousa, W. P. 1979. Disturbance in marine intertidal boulder fields: the nonequilibrium  maintenance of species diversity. Ecology 60:1225‐1239.       Appendix  1: Raw  invertebrate data  collected  from  47  streams  (Chapter 2)  in  the  Tongariro  National Park, New Zealand, between February and April 2007.   Values are mean number of  individuals  collected  from  five  Surber  samples  (area = 0.1 m2) at each  site.    Site names are  given in Appendix 2.    No.  Taxon  1 2 3 4 5  6  7 8 Ephemeroptera  1  Deleatidium spp.  56  43.6  71.2  80  17.4  23.8  51.6  28  2  Nesameletus ornatus  0  0.4  0.2  0  0  0  0  0  3  Coloburiscus humeralis 3.2  1.2  24.4  18.8  0.4  5.4  47  7.6  4  Acanthophlebia cruenata 0  0  0  0.2  0  0  0  0  5  Austroclima sepia  0  0  2.4  2.4  1.6  0.2  5.2  3  6  Austroclima jollyae  0  0.8  0  0  0  0.4  4  1  7  Mauiulus luma  0  0  0  0  0  0  0  0  8  Zephlebia spectabilis  0  0  0  0  0  0  0  0  9  Zephlebia borealis  0  0  0  0  0  0  0.6  0  10  Zephlebia versicolor  0  0  0  0  0  0  0  0  11  Zephlebia dentata  0  0  0  0  0  0  0  0  12  Zephlebia tuberculata  0  0.4  0  0  0  0  0  0  13  Amelotopsis perscitus  0  0  0  0  0.2  0  0  0  14  Neozephlebia scita  0  0  0  0  0  0  0.4  0  Plecoptera  15  Zealandoperla spp.  22.4  30.2  52.8  116  0.8  1  41.4  4.8  16  Zealandobius confusus gp. 0  0  0  0  0.2  0.2  0  0.4  17  Zealandobius furcillatus gp. 0  0  0  0  0  0  0  0  18  Austroperla cyrene  1.2  0.6  3.8  4  0  2.8  3  0.6  19  Stenoperla prasina  0  0.6  0.4  0.6  0  0  0  0  20  Taraperla pseudocyrene 0.2  0.2  0  0  0  0  0  0  21  Megaleptoperla dimunata 0  0  0  0  10.2  0  0  0  22  Megaleptoperla grandis 1.8  2.8  5.2  6.6  0  4.2  10  9.8  Trichoptera  23  Aoteapsyche colonica  0  1.4  9.2  16  0  3.4  55  0  24  Orthopsyche fimbriata  0.2  0  0  0  0  0  0  0  25  Orthopsyche thomasi  0.8  0  0  0  0  0  0.4  2.4  26  Early instar Hydrobiosidae 1.4  0.4  1.6  4  4.8  0  1.4  0  27  Psilochorema A  0.2  0.4  0.8  0  4.2  0.2  0.2  0  28  Hydrobiosis parumbripennis 0.2  0.2  1.2  0  1.6  0.2  0.2  1  29  Hydrobiosis silvicola  0  0  0  0  0  0  0  0  30  Hydrobiosis frater  0  0  0  0  0.4  0  0  0  31  Hydrobiosis spatulata  0.2  0  0  0  0  0  0  0  32  Hydrobiosis clavigera  0  0  0  0  0  0  0  0  33  Hydrobiosis charadrea  0  0  0  0  0  0  0  0  34  Neurochorema forsteri 0  0.2  0  0  0  0  0.4  0  35  Neurochorema confusum 0  0  0  0  0  0  0  0.2  36  Costachorema callistum 0  0  0.2  0  0  0  1  0.6  37  Costachorema xanthopterum 0  0  0  0  0  0  0.2  0  181           No.  9  11  12  13 14 15 16 17 18  19  20 21 1  3.6  2.8  16.4  52.2  143.6  72.6  145.6  106.4  9.6  38.4  29.6  16.8  2  0  0  0  0  0  0  0  4.2  0.8  3  2.6  1  3  0  0  1  0.4  0  4.8  12.8  9.4  46.4  1  4.8  0.2  4  0  0  0  0.8  0  0  1  0  0  0  0  0  5  0  0  2.4  0.2  0.4  0.2  0  1.2  101.6  2.4  1.6  0.8  6  0  0.8  0  0  0  0  0  0  0  0  0  0  7  0  0  0  0  0  0  0  0  0  0  0  0  8  0  0  0  0  0  0  0  0  0  0  0  0  9  0  0  0  0  0  0  0  0  0  0  0  0  10  0.8  0  0  0  0  0  0  0  0  0  0  0  11  0  0  0  0  0  0  0  0  0  2  0  0  12  0  0  0  0  0  0  0  0  0  0  0.2  0  13  0  0  0  0.2  0  0  0  0  0  0  0  0  14  0  0  0  0  0  0  0.2  0  0.8  0  0  0  15  3.4  14.8  15.6  14.8  12.4  25.8  42.4  19.6  8  10.4  18.6  73.2  16  0.6  0  0.8  0.8  23.2  0.2  0  0.4  0  2  0  0  17  0  0  0  0  0  0  0  0  0  0  0  0  18  0.4  0  0.2  0.8  7  11.8  0  11.6  10.6  0.8  0.8  0.8  19  0  0  0.4  0.4  1  0.2  1.8  0.8  0  1  0.8  0  20  0  0  0  0  0.4  0  0  0  4  0.6  0  0  21  0  0  0  0  0  0.2  0  0  0  0  0  0  22  1.2  0  24.2  2  21.2  4.2  0  5.2  0  0.8  1.2  0.8  23  29.6  16.2  0  0.2  0.4  1.4  11.2  9.6  20.8  1.8  0  5.6  24  0  0  0  0  0  0  0  0  0  0  2.4  0  25  0  0  0.8  0  0  0  0  0.2  0  0  0  0  26  0.8  0.8  0  4.4  2.8  0.4  3.2  1.6  12.8  0  7.2  6.6  27  0  0  0  0.2  2.4  0  4  2.4  4.8  5  2.6  0.8  28  1.2  0  2.4  0.2  0  0.4  0  1.6  16  0.6  1.8  2.8  29  0  0  0  0  0.2  0  0  0  0  0  0  0  30  0  0  0  0  0  0  0  0  0  0  0  0  31  0  0  0  0  0  0  0  0  0.8  0  0  0  32  0  0  0  0  0  0  0  0.2  0  0  0  0  33  0  0  0  0  0  0  0  0  0  0  0  0  34  0.8  0  0  1.2  0  0  0  0  0.8  0  0.6  1  35  0  0  0  0  0  0  0  0  0  0  0  0  36  1.8  0.4  0.8  0  0.6  0  0.2  0  0.8  0  1.6  0.8  37  0  0  0  0  0  0  0  0  0  0  0  0    182           No.  22  23  24  25 26 27 28 29 30  32  33  34 1  113.6  69.2  8.2  50.8  83.2  52.8  34.8  87.4  4.6  0  5.6  56  2  8  7.4  1.8  0.4  0  0  13.4  0  0  0  0  0  3  8.8  23  0  3.2  0  1  0  0  0  0  0  0  4  0  0.4  0  0  0  0  0  0  0  0  0  0  5  67.2  11.2  0.8  3.4  0.4  0  0  0  0  0  0  2.4  6  0  1  0  0  0  2.4  3.4  0  0  0  0  0  7  0  0  0  0  0  0  0  0  0  0  0  0  8  0  12.8  0  0  0  0  0  0  0  0  0  0  9  0  0  0  0  0  0  0  0  0  0  0  0  10  0  0  0  0  0  0  0  0  0  0  0  0  11  0  7.2  0  0  0  0  0  0  0  0  0  0  12  0  0  0  0  0  0  0  0  0  0  0  0  13  0  0.2  0  0  0  0  0  0  0  0  0  0  14  0  3.2  0  0  0  0  0  0  0  0  0  0  15  7.2  2.2  51.8  51  21.8  39.2  47.8  0  14.4  0.6  42.6  81.6  16  0  0.4  0  0  0  0  0.2  2.8  0  0  1  2.2  17  0  0  0  0  0  0  0  0  0  0  0  0  18  8  0.8  0  0.2  0  0.2  0  1.8  0  0  0.2  5.6  19  0  0.2  0  0  0  0  0  0  0  0  0  0  20  2.4  0  0  0  0  0  0  0  1.6  0  0  0  21  0  0  0  0  0  0  0  0  0  0  0  0  22  7.2  0.8  0  0  0.8  1.8  0.2  1.6  0  0  0.6  1.8  23  0.8  0.8  13.8  98.2  21.4  9  3.8  0  0  0  19.8  4  24  0  0  0  0  0  0  0  0  0  0  0  0.2  25  0  0  0  0  0  0  0  0  0  0  0  0  26  2.4  0  5.6  2.4  3  5.2  0.6  0  1  0.2  0.6  5  27  2.4  4.6  0  0.8  0.2  0  0  0  0.8  0  0  1  28  4  0.8  0.8  1.6  5.6  2.2  0.4  33  1.6  0  0.2  1.4  29  0  0  0  0  0  0  0  0  0  0  0  0  30  0  0  0  0  0  0  0  0  0  0  0  0  31  0  0  0  0  0  0  0  0  0  0  0.2  0  32  0  0  0  0  0  0  0  0  3.4  0  0  0  33  0  0  0  0  0  0  0.2  0  0  0  0  0  34  0  0  0.8  0.2  1.8  0.2  0  0  0  0  0  0  35  0  0  0  0  0  0  0  0  0  0  0  0  36  3.2  0  2.4  0  1.2  0  0.2  0  0  0  0  2.4  37  0  0  0  0.4  0  0  0  0  0  0  0  0.2    183           No.  35  36  37  38 39 40 41 42 43  44  45  46 1  76.8  153.6  15.8  8.4  49.8  15  25.2  32.8  72.8  73.2  7.4  6.4  2  0  0  0  0  0  0  0  0  0.2  4.8  0.2  0  3  13.6  24.8  0  0  20  0.8  1.2  14.6  22.8  22.6  1.6  0  4  0  0  0  0  0.2  0  0  0  0  0  0  0  5  116  37.6  1  0  1  1.4  0  62.4  9.4  3.2  12.2  0  6  0  0  0  0  1.8  0  11  0  0  0  0  0  7  0  0  0  0  2.2  0  0  0  0.2  0  0  0  8  0  0  0  0  0  0  0  0  0  1  0  0  9  0  0  0  0  0  0  0  0  0  0  0  0  10  0  0  0  0  0  0  0  0  0  0  0  0  11  8  0  0  0  0  0  0  0  0  0  4.8  0  12  0  0  0  0  0  0  0  1.6  0  0  0  0  13  0  0  0  0  0  0  0  0  0  0  0  0  14  0  0  0  0  0  0  0  0  0  0.2  0.2  0  15  0  35.2  8.4  19.4  18.4  29.4  7.6  21.6  14.2  44.2  0  19.6  16  150.4  20  2  4.4  0.8  12.6  0  4  19.2  1.8  2.4  0  17  0  0  0  0  0  0  1.6  0  3.6  0  0  0  18  13.6  0  0  0  1.6  0  3  2.4  7  3.8  1  0  19  0  0  0  0  1.4  0.2  0  0  1.8  0  1.6  0  20  0.8  0  0  0  0  0  0  3.4  0  0.2  0  0  21  0  0  1  0  0  0  0  0  0  0  0  0  22  14.4  3.2  0  2.2  0.6  0.2  0.8  0.8  1.8  0  0  0  23  0  0  0.4  1  2.2  6.2  13.6  0  0  2  0  1.8  24  32.8  0  0  0.4  0.4  0  1.6  23.4  30.6  0  6.8  0  25  0  12  0  0  0  0  0.8  0  0  0  0  0  26  8.8  2.4  0.2  1  3.2  3.2  7.8  4.8  14.2  2.4  0  0.6  27  0  0  0  0  0.4  0  0.8  0  1  3.8  1.4  0  28  7.2  0.8  0.4  2  6.2  1.4  3.4  0.2  2  1  0  0  29  0  0  0  0  0  0  0  0  0.6  0.2  0  0  30  0  0  0  0  0  0  0  0  0  0  0  0  31  0  0  0  0  0  0  0  1.4  0  0  0  0  32  0  0  0  0  0  0  0  0  0  0.4  0  0  33  0  0  0  0  0  0  0  0  0  0  0  0  34  3.2  0.2  0  0  1.2  3.2  1.8  0  0.8  0.2  0  0  35  0  0  0  0  0  0  0  0  0  0  0  0  36  0  0  0  0.4  1.4  0  2  1.6  3  1.2  0  0  37  0  0  0.2  0  1.2  0  0  0  0  0  0  0    184           No.  47  48  49  1  42.4  13.8  102.2  2  0  0.2  0  3  0  0  0  4  0  0  0  5  0  0.2  1  6  0  0  0  7  0  0  0  8  0  0  0  9  0  0  0  10  0  0  0  11  0  0  0  12  0  0  0  13  0  0  0.2  14  0  0  0  15  31.2  38  13.4  16  0.4  0  0  17  0  0  0  18  0.6  0.2  0.8  19  0.2  0  0  20  0  0  0  21  0.2  0  0.8  22  0  0  0  23  4.8  15.4  1  24  0  0  0  25  0  0  0  26  1.8  4  3.6  27  1  0.4  0  28  0  0  0.8  29  0  0  0  30  0  0  0  31  0  0  0  32  0  0  0  33  0  0  0  34  0  0  2.4  35  0  0  0  36  0.4  0  0.2  37  0  0.2  0  185     No.  Taxon 1 2 3 4 5  6  7 8 38  Costachorema psaropterum 0  0  0  0  0  0  0  0  39  Hydrobiosella mixta  0.2  0  0  0  0  0  0  0  40  Hydrochorema crassicaudatum 0  0  0  0  0  0  0  0  41  Pycnocentria evecta  0  0.2  1.2  1.6  0  0.2  5.6  0  42  Pycnocentria funerea 0.2  0  0  0  0  0  1.6  0.6  43  Pycnocentria sylvestris 0  0  0  0  0  0  0  0  44  Pycnocentrella eruensis 0  0  0  0  0  0  0  0  45  Pycnocentrodes aeris 0  6.8  9.2  0  0  0.2  43.6  0.4  46  Beraeoptera roria  2.2  48.4  75.2  24.8  0  5.2  154.4  34.4  47  Confluens hamiltoni  0  2.4  0  0.2  0  0  2  0.4  48  Oeconesus similis  0  0  0  0  0  0  0  0  49  Oecetis sp.  0  0  0  0  0  0  0  0  50  Hudsonema amabile 0  0  0  0  0  0.2  0  0  51  Hudsonema alienum 0  0  0  0  0  0  0  0  52  Oxyethira albiceps  1.2  0  3.6  0.8  0  0  2.8  0  53  Paroxyethira hintoni  0  0  0  0  0  0  0  0  54  Olinga feredayi  0  4.8  3.4  8  0  0.6  4  1  55  Helicopsyche poutini  0  1.8  1.4  7.2  0.4  0  1  0.4  56  Zelolessica cheira  0.8  0  0  0  3.2  0  0  4.8  57  Allocentrella sp.  0  0  0  0  0  0  0  0  58  Caddis Pupae  0  0  0  0  0  0  0.6  0  59  Polyplectropus sp.  0  0  0  0  0  0  0  0  60  Philorheithrus agilis  0  0  0  0  0  0  0  0  Diptera  61  Orthocladiinae A  0  0  0  0.8  0  0  0  0.4  62  Orthocladiinae B  0.2  0  1.2  4  0.6  1.2  0.6  0  63  Orthocladiinae C  2.4  0  0  7.2  3  1  25.6  0  64  Orthocladiinae D  4.2  1  0  0  0  0  0  0  65  Orthocladiinae E  8.4  1.8  0.2  0  84.6  0.4  19  3.2  66  Orthocladiinae F  0  0  0  0  1  0  0  0  67  Chironominae A  35.8  0.4  0  0  6.2  0.2  1.4  0.8  68  Maoridiamesa sp.  68.2  13  0.2  7.2  0.8  0.2  16.8  2.6  69  Tanypodinae A  0  0  0  0  0  0.2  0.2  0  70  Chironomidae pupae 7.4  0.2  0  0  4  0.2  2.8  0  71  Eriopterini (other)  1  0.8  0.6  1.6  2.8  3.2  0.4  1  72  Molophilus sp.  0.2  0  0  0  0  0  0  0  73  Aphrophila neozelandica 4  2  6.8  5  0  2.6  20.6  0  74  Austrosimulium spp.  0  0  1  1.6  3.6  0  0.4  0  75  Mischoderus sp.  0.2  0  0  0  0.2  0  0  0  76  Empididae  0.6  0  2  0  0.8  0  0.6  0  77  Limonia sp.  0  0  0  0  0  0  0  0  78  Neocurupira hudsoni gp. 0  0  0  0  0  1.2  0  0  79  Paradixa sp.  0  0  0  0  0  0  0  0      186     No.  9  11  12  13 14 15 16 17 18  19  20  21 38  0  0  0  0  0  0  0  0  0  0  0  0  39  0  0  0  0  0.2  0  0.2  0  0  0.4  0  0  40  0  0  0  0  0  0  0  0  0  0  0  0  41  0  0  0  0  0  0  0  0  24.8  0.2  0.2  0  42  0  0  0  0.4  0.2  0  0  0  0  9.2  0  0  43  0  0  0  0  0  0  0  0  0  0  0  0  44  0  0  0  0  0  0  0  0  0  0  0  0  45  0.8  0  6.4  0  0  0  0  0.4  4  0  0  0  46  0  0  0.2  6.4  0.2  35  16.8  18  0.8  0.2  0.8  1.6  47  0.2  0  0  0  0  0  0  0  36.8  0.6  0  1.6  48  0  0  0  0  0.2  0  0  0  0  4.2  0  0  49  0  0  0  0  0  0  0  0  0  2.8  0  0  50  0  0  0  0  0  0  0  0  0  0  0  0  51  0  0  0  0  0  0  0  0  0  0  0  0  52  0  0  1  0  3.6  0  0  0  1.6  0  1.6  0.8  53  0  0  0  0  0  0  0  0  0  0.2  0.8  0  54  0.4  0  1.4  1.8  0  6.4  7.2  3.2  0  0.2  4.8  1.6  55  0  0  0.8  0  0  21.6  0.8  5.6  0  6.8  0  0  56  2.6  0  4.2  1.4  1.8  0  0  0  76  0.6  0  0  57  0  0  0  0  0  0  0  0  24  0.2  0  0  58  0  0  0  0  0  0  0  0  0  0  0  0  59  0  0  0  0  0  0  0  0  0  0.4  0.2  0  60  0  0  0  0  0  0  0  0  0  0  0  0  61  0  0  0  0  0.4  0  7.2  0.8  0  4.4  0  0  62  10.4  5.6  6  25.2  7.6  1.6  0.8  0.8  8.8  2.6  0  0  63  6.8  0  1.6  10  0  0  0.8  0  5.6  2.2  541.6  170.8  64  0  0  0  9.2  16.6  0  0  0  0  0  0  0  65  11.6  2  0  34  33  3.2  0  2  72.8  12.6  57.6  67.6  66  0  0  0  0  0.4  0  0  0  0  0  4  0  67  10.4  1.6  8  5  35.4  0  0  0.8  0.8  0  0.8  0  68  82  2.6  87  11.8  4.8  0  0  0  6.4  0.8  132  81.2  69  0  0  0  0  0  0  0  0  0  0.2  0  0  70  0  0  1  0.4  6.2  0  0  0  1.6  0  16.8  9.2  71  0  2.2  0  1.2  2.6  2.6  0  0.8  0  0  0  0  72  0  0  0  0  0  0  0  0  0  0  0  0  73  0.6  1.2  9.2  5.4  5.4  2.4  7.2  6.6  8  0  0.8  1.6  74  0  2.4  2.4  0  0  0  0  0  4  0  1.6  0  75  0  0  0  0  0  0  0  0  0  0  0  0  76  0.2  1.4  8.4  0  0.2  0  0  0.8  0  0  0  0  77  0  0  0  0  0  0  0  0  0  0  0.8  0  78  0  0  0  0  0  1.8  0  0  0  0  0  0.2  79  0  0  0  0  0  0  0  0  0  0.4  1.6  0      187     No.  22  23  24  25 26 27 28 29 30  32  33  34 38  0  0  0  0  0  0  0  0  0  0  0  0  39  0  1.6  0  0  0  0  0  0  0  0  0  0  40  0  0  0  0  0  0  0  0  0  0  0  0  41  0  0  0  0.2  0  0  0  0  0  0  0  0  42  9.6  5.2  0  0  0  1.6  0  0  0  0  0  1.8  43  0  0  0  0  0  0  0  0  0  0  0  0  44  0  0  0  0  0  0  0  0  0  0  0  0  45  30.4  10.8  0  0  0.4  13.6  1.4  0  1.6  0  0.2  0.8  46  6.4  0  6.2  29  2.6  29.6  0.6  0  0  0  6.6  2.4  47  0  0  2.4  0  27.8  10.4  0  0  0  0  0  6.4  48  0  0  0  0  0  0  0  0  0.2  0  0.2  0  49  0  0  0  0  0  0  0  0  0  0  0  0  50  0  0  0  0  0  0  0  0  0  0  0  0  51  0  0  0  0  0  0  0  0  0  0  0  0  52  4  0  11.4  0.2  1  0.4  0  0  0.8  0  0.4  0  53  0  0  0  0  0  0  0  0  0  0  0  0  54  0  1.8  0  1  0  0  0  0  0  0  0  0  55  2.4  43.6  0  0  0  0  0  0  0  0  0  0  56  0.8  0  0  0  0.4  0  0.6  181.4  7.2  0.2  0.2  1.6  57  0  0.8  0  0  0  0  0  0  0  0  0  0  58  0  0  0  0  0  0  0  0  0  0  0.4  0  59  0  0  0  0  0  0  0  0  0  0  0  0  60  0  0  0  0  0  0  0  0  0  0  0  0  61  0  2.4  0  0  0  1.6  0.2  0  0  0.2  0  0  62  0  0.8  15.2  0  0  29.6  1.2  157.4  103.2  2  6.2  4.2  63  0  0.8  133.6  3.6  0  17.6  1.8  0.4  8  0  1.6  2.6  64  0  0  16.8  0  57  4.8  0  0  0  0  0.6  1.2  65  23.2  0  41.6  2  0  16.8  0  0  20  0  6.2  1.6  66  0  0  0  0  0  0  0.2  0  0  0  0  0  67  2.4  0  0  0  0  3.2  0.4  0.4  0  0  0  0  68  16.8  0.8  535.2  15.6  20.2  92  0.2  692.8  189.6  0  19  0.6  69  0  0.8  0  0.2  0  0  0  0  0.2  0  0  0  70  0  0  5.6  0.2  1.8  0.8  0.2  9.4  2.4  0.6  0.4  0.4  71  2.4  0  0  1.6  0.6  0.2  1.4  0  0  0  0.8  4  72  0  0  0  0  0  0  0  0  0  0  0  0  73  2.4  1  1.6  6.4  18.2  14.4  3  0  0  0  0  3.6  74  3.2  0  2.4  0  0  0.8  2  0  1.6  0  4.6  1.2  75  0  0  0  0  0  0  0  0.2  0  0  0.2  0  76  0  0  0  0  0  0  0  20  0.8  0  0  0.8  77  0  0  0  0  0  0  0  0  0.8  0  0  0  78  0  0  0  0  0  0  0  0  0  0  0  0  79  0  0.8  0  0  0  0  0  0  0  0  0  0      188     No.  35  36  37  38 39 40 41 42 43  44  45  46 38  0  0  0  0  0.8  0  0  0  0  0  0  0  39  0.8  0  0  0  0  0  0  0  0  0  0.8  0  40  0.2  0  0.2  0  0.2  0  0  0.4  0.2  0  1  0  41  28  0  2.4  0  0.6  0.6  18.8  0  0  0  0  0  42  0  4  0  0  0  0.6  17  22.4  3.8  0.2  0  0  43  0  0  1  0  0  0  0  0  0  0  0  0  44  0  0  0  0  0  0  0  0  0  0  1.6  0  45  0  0  11.2  2  0.2  0.2  8.6  0  0  0  0  0  46  0  0.8  14.6  2.2  7.8  0.8  1.2  0  0  30.4  0  0  47  0  0  0  0  7.4  0  5  0  0  0.4  0  0  48  0  0  0  0  0  0  0  0  0  0  0  0  49  0  0  0  0  0  0  0  0  0  0  0  0  50  0  0  0  0  0  0  0  0  0  0  0  0  51  0  0  0  0  0  0  0.2  0  0  0.2  0  0  52  1.6  0.8  0  0  0.4  0  0  0  0.2  0  0  0  53  0  0  0  0  0  0  0  0  0  0  0  0  54  0  0  0  0  0.4  0.2  0  0  0  1  0  0  55  10.4  0  0  0  0  0  0.8  3.2  0.6  22.6  7.6  0  56  19.2  1.6  25  0.2  15.2  0  27  12.8  2.2  0  0  0  57  8  0  0  0  0  0  0.8  2.4  0  0  2  0  58  0  0  0  0  0  0.4  0  0  0.2  0  0  0  59  0  0  0  0  0.2  0  0  0  0  0  0  0  60  0  0  0  0  0  0  0  0  0  0  0.2  0  61  0  0  0  0  1  0  0.2  0.8  0  0  1.2  0  62  0  0  34  38  58.8  39.2  3  4  27.8  5.4  0  0  63  7.2  4  1.6  0  18  11  0  1.6  0  0  0  0  64  0  0  0  0  4.2  1.8  0  0  9.6  0  0  0  65  32  3.2  12.4  2.2  0  3.8  2.4  4  0  0  0.6  0  66  0  0  0  0  0  0  0  0  0  0  0  0  67  6.4  0  0  0  1.2  4  0  0.8  0.2  0  0  0  68  20.8  4  29.4  50.2  4.2  37  20.6  0  0  0.8  0  0  69  0  0  0  0  0  0  0  0  0  0  0  0  70  0  0.8  3.6  0  6.8  3.6  0.8  0  1.4  0.4  0  0  71  0.2  0.8  0.2  0.6  2  4.4  0  0  1.6  0.4  0.2  1.4  72  0  0  0  0  0  0  0  1.6  0  0  0  0  73  6.4  0  0  0  20.2  2.4  1.6  0  3.6  5.8  0  0  74  4  0  0.8  1.4  0  0.6  0  3.2  1  0  0  0  75  0  0  0  0  0.6  0  0  0  0  0  0  0  76  0  0  0  0  0.4  0  0.4  0  0  0  0  0  77  0  0.8  0.4  0  0  0  1  0  0  0  0  0  78  0  0  0  0  0  0  0  0  0  0  0  0  79  0  0  0  0  0  0  0  0  0  0  0  0      189     No.  47  48  49  38  0  0  0  39  0  0  0  40  0.2  0  0  41  0  0  0  42  0  0  0  43  0  0  0  44  0  0  0  45  0  0  0.8  46  0  0  0.2  47  0  0  0  48  0  0  0  49  0  0  0  50  0  0  0  51  0  0  0  52  0.4  0  0.8  53  0  0  0  54  0.6  0  0.8  55  0  0  0  56  0  0  0.8  57  0  0  0  58  0  0  0  59  0  0  0  60  0  0  0  61  0  0  0  62  9  8  27.4  63  4.8  25.2  74.8  64  0  0  10.4  65  0.4  1  0  66  0  0  0  67  0  0.4  5  68  3  1.2  7.8  69  0  0  0.8  70  3.2  1  2  71  3.2  0.6  6.2  72  0  0  0  73  0.2  0.4  7.4  74  1.8  2.8  0.8  75  0  0  0  76  0  0  1  77  0  0  0  78  0  0  0  79  0  0  0      190     No.  Taxon  1 2 3 4 5  6  7 8 80  Nothodixa sp.  0  0  0.2  0  0.4  0  0  0  81  Ephydrella thermarum 0  0  0  0  0  0  0  0  82  Ceratopogonidae  0  0  0  0  0  0  0  0  83  Muscidae  0  0.8  0  0  0.8  0  0  0  Coleoptera  84  Elmidae Hydora spp.  9.2  13.6  24  20  0  2.4  2.6  12.4  85  Hydraenidae  0  2  2.4  0.8  1  0  1.4  0  86  Ptilodactylidae  0  0  0.2  0  0.2  0  0  0  87  Hydrophylidae  0  0.2  0  0  0.2  0  0  0  88  Scirtidae  0  0  0  0  0  0  0  0  Mollusca  89  Potamopyrgus antipodarum 0  0  0  0.8  0  0  0.6  0  90  Latia neritoides  0  0  0  0  0  0  0  0  91  Gyraulus corinna  0  0  0  0  0  0  0  0  Other  92  Archichaulioides diversus 0  1  2.2  1.6  0.6  1.4  4.4  1.4  93  Platyhelminthes  0  0  0  0  0  0  0.4  0  94  Oligochaeta  1.8  0  2.8  2.4  27.2  0.2  0.4  0  95  Neuroptera Kempynus spp. 0  0  0  0  0  0  0  0  96  Nematoda  0  0.2  0  0  0  0  0  0  97  Ostracoda  0  0  0  0  0  0  0  0  98  Paricalliope spp.  0  0  0  0  0  0  0  0  99  Anisops assimilis  0  0  0  0  0  0  0  0  100  Nematomorpha  0  0  0  0  0.2  0  0  0        191     No.  9  11  12  13 14 15 16 17 18  19  20  21 80  0  0  0  0  0  0  0  0  0  0  0  0  81  0  0  0  0  0  0  0  0  0  0  0  0  82  0  0  0  0  0  0  0  0  2.4  0  0  0  83  1  0  0  1.4  0  0  0  0  0.8  0  0.2  3.4  84  8.8  0  0.8  22  9.8  30.4  5.6  10  46  0.6  7.2  5  85  0  0  0  1.6  6  1.6  0.8  2.6  30.4  1.6  2.4  0  86  0  0  0  0  0.2  0  0  0  3.2  0.2  0  0  87  0  0  0  0  0  0  0  0.8  0.8  0  0  0  88  0  0  0  0  0  0  0  0  0  0  0  0  89  0  0  0.2  0  0  0  0  0  0  0.4  0  0  90  0  0  0  0  0  0  0  0  0  0  0  0  91  0  0  0  0  0  0  0  0  0  0  0  0  92  0  0  0  2.6  0.4  1.6  4.8  0.4  4  0  1.6  0.8  93  0  0  1  0  0  0  0  0  1.6  0.8  0  0  94  0  0.2  0.8  1  1.2  0  19.2  5.6  9.6  21.2  4  2  95  0  0  0  0  0  0  0  0  0  0  0  0  96  0  0  0  0.2  0  0  0  0  0  0.4  0  0  97  0  0  2.4  0  0  0  0  0  0  0  0  0  98  0  0  0  0  0  0  0  0  0  0  0  0  99  0  0  0  0  0  0  0  0  0  0  0  0  100  0.2  0  0  0  0  0  0  0  0  0  0  0      192     No.  22  23  24  25 26 27 28 29 30  32  33  34 80  0  0  0  0  0  0  0  0  0  0  0  0  81  0  0  0.2  0  0  0  0  0  0  0  0  0  82  0  0  0  0  0  0  0  0  0  0  0  0  83  0  0  3  0.8  0.8  9.6  0.2  2  16.6  0  0  0  84  80  8.8  2.4  15.6  24.4  20.8  20  6.8  12.8  0  2.8  22.8  85  16  5.2  0  0.8  0  0  0  0  0  0  0  0.4  86  1  0  0  0  0  0  0  0  0  0  0  0.8  87  0  0  0  0  0  0  0  0  0  0  0  0  88  0  0  0  0  0  0  0  0  0.2  0  0  0  89  0  4.6  0  0  0  0  0  0.2  0.8  0  0  0  90  0  6.6  0  0  0  0  0  0  0  0  0  0  91  0  0  0  0  0  0  0  0  0  0  0  0  92  1.6  3.4  0  0  0  0  0  0  0  0  0  1.6  93  6.4  4.4  0  0  0  0  0.4  6.4  0  0  0.4  0  94  17.6  1.8  12  4  1.6  0.8  0  1  59.2  0  0  5.8  95  0  0  0  0  0  0  0  0  0  0  0  0  96  0  0.4  0.8  0  0  0  0  0  0.8  0  0  0  97  0  0  0  0  0  0.8  0  0  0  0  0  0  98  0  0  0  0  0  0  0  0  0  0  0  0  99  0  0  0  0  0  0  0  0  0.2  0  0  0  100  0  0  0  0  0  0  0  0  0  0  0  0      193     No.  35  36  37  38 39 40 41 42 43  44  45  46 80  0  0  0  0  0  0  0  0  0  0  0.2  0  81  0  0  0  0  0  0  0  0  0  0  0  0  82  0  0  0  0  0  0  0  0  0  0  0  0  83  0  0  0.4  0.2  0.6  1.2  3  0  0  0  0  0  84  28  120  0.2  0  10.6  11.4  34.4  1.6  37  4.4  0  1.2  85  22.4  0.8  0  0  0.2  0  0.8  0  3.8  7.8  0  0  86  14.4  0  0  0  0  0  1  6.4  1.4  0  0.2  0  87  0  0  0  0  0  0  0  0  0  0  0  0  88  0  0  0  0  0  0  0  0  0.8  0  0  0  89  8  0  0  0  0  0  0  0  0  0.6  0.6  0  90  0  0  0  0  0  0  0  0  0  0  0  0  91  0.2  0  0  0  0  0  0  0  0  0  0  0  92  0  4.2  0  0  4  0  5.4  0  0.6  6  0  0  93  7.2  0  1.8  1.6  0.2  0  4  2  0  0.4  2.4  0  94  14.4  6.4  25  1.6  23.2  4.8  16.2  3.2  3  1.2  0.2  0.2  95  0  0  0  0  0  0  0  0  0  0.2  0  0  96  0  0  0.6  0.6  0  0  0  0  0  0  0  0  97  0  0  0  0  0  0  0  0  0  0  0  0  98  0  0  0  0  0  0  0  0.8  0  0  14.8  0  99  0  0  0  0  0  0  0  0  0  0  0  0  100  0  0  0  0  0.4  0  0  0  0  0  0  0      194     No.  47  48  49  80  0  0  0  81  0  0  0  82  0  0  0  83  0  0  0.8  84  4.2  2.4  20.6  85  1  0  0  86  0  0  0  87  0  0  0  88  0  0  0  89  0  0  0  90  0  0  0  91  0  0  0  92  0  0  0  93  0  0  0  94  0.4  3.8  0  95  0  0  0  96  0  0  0  97  0  0  0  98  0  0  0  99  0  0  0  100  0  0  0      195     Appendix  2:  Sampling  location  information  for  the  47  streams  used  in  Chapter  2  in  the  Tongariro National Park, New Zealand, sampled between February and April 2007.  Altitude =  m asl.  * Indicates the 24 New Zealand sites used in Chapter 5.    Site no.  Site name  Easting Northing  Altitude 1  Te Whaiau stream 2735629 6236670  643 2*  Whanganui @ Te Porere 2734447 6236741  654 3*  Whanganui d/s intake 2735280 6239631  604 4*  Mangatepopo by camp 2731856 6234203  752 5  Rotoaira trib  2745962 6232119  570 6  Mangahouhounui  2751536 6231933  540 7*  Poutu Stream  2752035 6232647  518 8  Puketerata SH1  2752750 6228950  590 9*  Oturere Stream SH1 2748265 6221043  809 11*  Mangatoetoenui @ SH1 2745841 6215290  971 12*  Unnamed @ SH1  2748253 6219911  885 13  Waihaha stream  2752693 6222742  651 14  Trib of Mangawhero on Turoa rd 2723107 6202308  912 15  Mangawhero @ Turoa rd 2719503 6198649  663 16  Mangaeteroa Stream @ Horopito 2714942 6203708  736 17  Makotuku River @ Horopito 2715050 6203696  713 18  Orautoha stream @ Middle rd 2713091 6205484  712 19  Taurewa stream d/s intake 2730850 6235993  621 20*  Tauwhitikuri Stream d/s intake 2730873 6236020  621 21*  Mangatepopo d/s dam 2730962 6236051  624 22  Tauwhitikuri Stream u/s intake 2731045 6233786  728 23  Okupata d/s intake 2728772 6235536  694 24*  Whakapapa d/s intake 2723280 6228877  680 25  Whakapapa u/s intake 2723465 6228824  697 26*  Whakapapanui @ SH4 2726914 6225525  835 27*  Whakapapaiti @ SH4 2723631 6222537  859 28*  Trib of Ohinepango @ old hut 2740786 6218482  1106 29*  Ohinepango Springs 2741269 6217420  1091 30*  Waihohonu Springs 2740293 6220717  1158 32*  Wahianoa stream u/s intake 2737514 6199145  934 33*  Unnamed Karioi forest stream 2736402 6199318  935 34*  Te Unuunuakapuateariki stream 2730309 6192788  701 35  Waiharakeke stream 2733203 6192405  735 36  Unnamed Karioi forest stream 2732704 6193961  725 37  Unnamed Karioi forest stream 2733009 6194725  749 38  Unnamed Karioi forest stream 2740109 6198462  914 39  Mangatawai @ Tongariro confluence  2750705 6221184  660 40*  Tongariro d/s Poutu intake 2754050 6226836  456 41  Waitaki stream d/s Rotokura lake 2726571 6192156  627 42  Waiharuru trib  2725422 6196222  710 43  Waiharuru stream 2725000 6196222  710 196           Site no.  Site name  Easting Northing  Altitude 44*  Makomiko stream 2716563 6216130  746 45  Unmarked spring  2725250 6198400  770 46*  Mangaturuturu river 2716300 6208300  820 47*  Waipakihi @ end of road 2749450 6216500  857 48*  Tongariro d/s Rangipo dam 2750083 6218070  813 49*  Tongariro @ pillars of hercules 2750900 6220800  660     197     Appendix 3: Raw invertebrate data from 16 study sites (Chapter 3) in Tongariro National Park  and Hawke’s Bay  of  the North  Island  of New  Zealand  collected  on  four  occasions  between  February  2008  and  July  2009.    Values  are mean  number  of  individuals  collected  from  five  Surber samples (area = 0.1 m2) at each site on each occasion.  Site names are give in Table 1 in  Chapter 3.   0208 = February 2008, 1008 = October 2008, 0209 = February 2009, 0709 =  July  2009.    No.  Taxon N P1 .0 20 8  N P2 .0 20 8  N P3 .0 20 8  N P4 .0 20 8  N P5 .0 20 8  N P6 .0 20 8  N P7 .0 20 8  N P8 .0 20 8  Ephemeroptera  1  Deleatidium spp.  0  11.2  70.4  94.8  3.6  5  19.6  96.8  2  Nesameletus ornatus  0  0  1.4  0  0  0  0  0  3  Coloburiscus humeralis 0  0  26.4  0.6  0  0  0  100.4  4  Acanthophlebia cruenata 0  0  0.2  0  0  0  0  0  5  Austroclima sepia  0  0  29.2  0  4.2  0  0  8.2  6  Austroclima jollyae  0  0  0  3  0  2  4  7.6  7  Mauiulus luma  0  0  0  0  0  0  0  0  8  Zephlebia spectabilis  0  0  2  0  0  0  0  0  9  Zephlebia versicolor  0  0  0.2  0  0  0  0  0  10  Zephlebia dentata  0  0  0  0  0  0  0  0.2  11  Zephlebia tuberculata  0  0  0  0  0.2  0  0  0  12  Amelotopsis perscitus  0  0  0  0  0  0  0  0  13  Neozephlebia scita  0  0  4  0  0  0  0  0.4  Plecoptera  14  Zealandoperla spp.  0  31.4  8.8  25.8  2.6  4.4  14.2  71.8  15  Zealandobius confusus gp. 0  0.6  0  0.4  2.2  0  0  1.2  16  Austroperla cyrene  0  0.6  1.6  1  0  0  0.2  7  17  Stenoperla prasina  0  0  0.4  0  0  0  0  0  18  Taraperla pseudocyrene 0  0.2  2.8  0  0  0  0  0  19  Megaleptoperla grandis 0  5.8  0.2  3.6  0  1.6  0.6  6.6  20  Megaleptoperla dimunata 0  0  0  0  0  0  1.8  0  Trichoptera  21  Aoteapsyche colonica  0  22  3.2  7.8  0  13.6  11.6  172.8  22  Orthopsyche fimbriata 0.2  0  0  0  0  0  0  1  23  Orthopsyche thomasi  0  0  0.2  0  0  0  0  0  24  Early instar Hydrobiosidae 0  3.4  8  4.6  5.6  4  1.8  0  25  Psilochorema A  0  0  2.2  1.2  0  0  0  0  26  Psilochorema B  0  0  0  0  0  0  0  1.2  27  Psilochorema C  0  0  0  0  0  0  0  0.4  28  Psilochorema D  0  0  0  0  0  0  0  0  29  Psilochorema E  0  0  0  0  0  0  0  0  30  Hydrobiosis parumbripennis 0.4  1  4  1.8  2.4  2  3.4  2  31  Hydrobiosis umbripennis 0  0  0  0  0  0  0  0      198               No.  H B1 .0 20 8  H B2 .0 20 8  H B3 .0 20 8  H B4 .0 20 8  H B5 .0 20 8  H B6 .0 20 8  H B7 .0 20 8  H B8 .0 20 8  N P1 .1 00 8  N P2 .1 00 8  N P3 .1 00 8  N P4 .1 00 8  1  76.4  368  49.6  22.4  109.6  128  1.2  8  1.6  10  38  66  2  0  0  0  0  0  0  0  0  0  0  0.4  3.2  3  0.6  0  0.4  0  1.2  0.4  0  0  0  1.2  60.4  2.8  4  0  0  0  0  0  0  0  0  0  0  0  0  5  0.8  0  13.6  11.2  4.4  10.4  0  17.2  3.44  0  31.6  0.8  6  0  0  0  0  0  0  0  0  0  0  0  0  7  0  0  1.6  0  0  0  0  0  0  0  0  0  8  0  0  0  0  0  0  0  0  0  0  0  0  9  0  0  0  0  0  0  0  0  0  0  0  0  10  0  0  7.2  4.8  0  1.6  0  1.8  0.36  0  0  0  11  0  0  0  0  0  0  0  0  0  0  0  0  12  0  0  0  0  0  0  0  0  0  0  0  0  13  0  0  0  0  0  0  0  0  0  0  4.8  0  14  0  0  0  0  0  0  0  0  0  3  16  1.8  15  0  0  0  0  0  0  0  0  0.2  0.2  0  0  16  0  0  0  0  0  0  0  0  0  0.6  0.8  1.6  17  0  0  0  0  0  0  0  0  0  0  0.8  0.8  18  0  0  0  0  0  0  0  0  0  0  0.4  0  19  0  0  0  0  0  0  0  0  0  2.6  0  3.2  20  0  0  0  0  0  0  0  0  0  0  0  0  21  8.8  8.8  112.8  32.8  199.2  483.2  6.8  9  1.8  1.6  0.2  2  22  0  0  0  0  0  0  0  0  0  0  0  0  23  0  0  0  0  0  0  0  0  0.2  0.6  0  0  24  0  0  0  0  1.8  0  4.4  0.8  0.96  0  5.2  0.8  25  9.2  6.4  3.4  2.4  4.4  8  0  0  0  0  0  0  26  0  0  0  0  0  0  0  5.4  1.08  0  0  0  27  0  0  0  0  0  0  0  0  0  0  0  0  28  0  0  0  0  0  0  0  0  0  0  0  0.8  29  0  0  0  0  0  0  0  0  0  0  0  0  30  3.2  7.2  1.6  8  4.2  4  5.8  0.8  0.16  0  0  0  31  0  0  0  0  0  0  0  1  0.2  0  0  0      199               No.  N P5 .1 00 8  N P6 .1 00 8  N P7 .1 00 8  N P8 .1 00 8  H B1 .1 00 8  H B2 .1 00 8  H B3 .1 00 8  H B4 .1 00 8  H B5 .1 00 8  H B6 .1 00 8  H B7 .1 00 8  H B8 .1 00 8  1  1.4  1  53  12.4  87.6  165.2  145.2  9.4  86.8  84.8  8  30.8  2  0  0  1  0  0  0  0  0  0  0  0  0  3  0  0  0.4  6.8  0  0  2.4  0  0.8  0.2  0  0  4  0  0  0  0  0  0  0  0  0  0  0  0  5  0  0  0  1.6  0  0  26.6  2  0  0.8  21.6  6  6  0  0  0.2  0  0  0  0  0  0  0  0  0  7  0  0  0  0  0  0  0  0  0  0  0  0  8  0  0  0  0  0  0  0  0  0  0  0  0  9  0  0  0  0  0  0  0  0  0  0  0  0  10  0  0  0  0  0  0  0  0  0  0  0  0  11  0  0  0  0  0  0  3.2  0  0  0  0.8  2.8  12  0  0  0  0  0  0  0  0  0  0  0  0  13  0  0  0  0  0  0  3.6  0  0  0  0  0  14  0.2  2.4  0.2  3.2  0  0.4  0  0  0  0  0  0  15  0  0  0.2  0.2  2.8  1.2  0.8  0  7.6  0.4  0  0  16  0  0  0  0.2  0  0  0  0  0  0  0  0  17  0  0  0  0  0  0  0  0  0  0  0  0  18  0  0  0  0  0  0  0  0  0  0  0  0  19  0  0.2  0.2  0  0  0  0  0  0  0  0  0  20  0  0  4.8  0  0  0  0  0  0  0  0  0  21  0.2  1.6  1  1  0.4  0  20  1.8  2  1.6  9.8  1.6  22  0  0  0.4  1.2  0  0  0  0  0  0  0  0  23  0  0  0  0  0  0  0  0  0  0  0  0  24  0  0  3.4  0  0  0  0  0.6  0  2.4  1.8  0  25  0  0  0  0  2  0.8  0  0.2  1.2  1.2  1.4  0  26  0  0  0  0  0  0  0  0  7.2  0  0  0  27  0  0  0.2  0  0  0  0  0  0  0  0  0  28  0  0  0.8  0  0  0  0  0  0.8  0  0  0  29  0  0  0  0  0  0  0  0  0  0  6.4  3.2  30  0.2  0.2  0  0.2  0  0.4  0.4  0.2  0  2.2  4  0  31  0  0  0  0  0  0  0  0  0  0  0  0      200               No.  N P1 .0 20 9  N P2 .0 20 9  N P3 .0 20 9  N P4 .0 20 9  N P5 .0 20 9  N P6 .0 20 9  N P7 .0 20 9  N P8 .0 20 9  H B1 .0 20 9  H B2 .0 20 9  H B3 .0 20 9  H B4 .0 20 9  1  0.2  6.2  29.8  31.8  5.6  0.2  54.2  43.8  192.4  62  173.6  46  2  0  0  0.2  0  0  0  0  1.8  0.4  0.2  0  0  3  0.2  0.6  25.6  0.4  0  0  0.8  63.8  0.8  0  2.6  0  4  0  0  0  0  0  0  0  0  0  0  0  0  5  0  1  49.6  0.2  0.2  0  0.2  3.8  0  0  55.2  8.6  6  0  0  0  0  0  0  0  0.8  0  0  0  0  7  0  0  2.4  0  0  0  0  0  0  0  0  0  8  0  0  3  0  0  0  0  0  0  0  0  0  9  0  0  0  0  0  0  0  0  0  0  0  0  10  0  0  0  0  0  0  0  0  0  0  4.8  3  11  0  0  0  0  0  0  0  0  0  0  0  0  12  0  0  0.2  0  0  0  0  0  0  0  0  0  13  0  0  5.6  0  0  0  0  0  0  0  4  0  14  0.2  24.4  6  16.4  3.2  0.4  12.8  26.4  1.4  0  0  0  15  0  0.2  0.8  0.2  1  0.2  0  0.8  0  0  0  0  16  0  2.4  2.2  0  0  0  0.2  6.4  0  0  0  0  17  0  0  0.2  0  0  0  0  0  0  0  0  0  18  0  0  0.2  0  0  0  0  0  0  0  0  0  19  0  1  0  0  0  0  1.2  7.4  0  0  0  0  20  0  0  0  0  0  0  0  0  0  0  0  0  21  0  14.6  0.8  0.2  0  0.6  3  67.8  5  6.4  363.2  14.4  22  0.4  0  0  0  0  0  0  0  0  0  0  0  23  0  0  0  0  0  0  0  0  0  0  0  0  24  0.4  0  12  0.8  0  0.2  3.4  6  0.4  2.6  0.8  0.8  25  0  0  3.6  1  0  0  0.8  0  3  2  3  0  26  0  0  0  0  0  0  0  0  0  0  0  0  27  0  0  0  0  0  0  0  0  0  0  0  0  28  0  0  0  0  0  0  0  0  0  0  0  0  29  0  0  0  0  0  0  0  0  0  0  0  0  30  0  0.6  0  0  11.2  0  1.8  5.6  2.6  1.2  4  3  31  0  0  0  0  0  0  0  0  0  0  0  0      201               No.  H B5 .0 20 9  H B6 .0 20 9  H B7 .0 20 9  H B8 .0 20 9  N P1 .0 70 9  N P2 .0 70 9  N P3 .0 70 9  N P4 .0 70 9  N P5 .0 70 9  N P6 .0 70 9  N P7 .0 70 9  N P8 .0 70 9  1  134.8  276.4  0  8.6  0.4  9.2  11.4  44.4  4  1.8  108.2  81.2  2  0.2  0  0  0.2  0  0.4  0.2  0.8  0  0  0.8  2.8  3  1.8  0  0  0  0  0.8  8  0.2  0  0  8.8  44.2  4  0  0  0  0  0  0  0  0  0  0  0  0  5  3.8  12.4  0  12.8  0  0  12  0.2  6.4  0  0  5.4  6  0  0  0  0  0  1  0  0  0  0  1  0.2  7  0  0  0  0  0  0  0  0  0  0  0  0  8  0  0  0  0  0  0  0  0  0  0  0  0  9  0  0  0  0  0  0  0  0  0  0  0  0  10  0  0.2  0  0.8  0  0  0  0  0  0  0  0.2  11  0  0  0  6.2  0  0  0  0  0  0  0  0  12  0  0  0  0  0  0  0  0  0  0  0  0  13  0  0  0  0  0  0  1.4  0  0  0  0  1  14  0.2  0  0  0  0.2  24  2  2.8  0.6  5  3  11  15  0  0  0  0  0  0.4  0.4  0.2  29.2  0  3.4  3.8  16  0  0  0  0  0  0.6  0.2  0  0  0  6  2  17  0  0  0  0  0  0  0  0  0  0  0  0  18  0  0  0  0  0  0  0  0  0  0  0  0  19  0  0  0  0  0  0.2  0  0  0  0  2.8  1.2  20  0  0  0  0  0  0  0  0  0  0  0  0  21  50.2  113.8  0  1.6  0.2  2.4  0  0  0  0  3.4  19.2  22  0  0  0  0  0  0.2  0  0  0  0  0  0  23  0  0  0  0  0  0  0  0  0  0  0  0.6  24  1  0.8  2.4  0.6  1.6  1  2.2  2  4.6  2.2  11.8  5.4  25  0  6  0  6.2  0  0.2  0.2  0.2  0.4  0  5.8  0.6  26  1.8  1.2  0  0  0  0  0  0  0  0  0  0  27  0  0  0  0  0  0  0  0  0  0  0  0  28  0  0  0  0  0  0  0  0  0  0  0  0  29  0  0  0.8  2.4  0  0  0  0  0  0  0  0  30  3  12.6  22.4  0  0  0.8  0  2.6  0.4  0.2  2.8  3.4  31  0  0  0  6  0  0  0  0  0  0  0  0      202               No.  H B1 .0 70 9  H B2 .0 70 9  H B3 .0 70 9  H B4 .0 70 9  H B5 .0 70 9  H B6 .0 70 9  H B7 .0 70 9  1  18.8  52.6  110.2  37.4  79.6  45.2  4.4  2  0  0  0  0  0.2  0  0  3  0  0  1.4  0  1.6  0  0  4  0  0  0  0  0  0  0  5  0  0  37.8  3.8  0  0  5  6  0  0  0  0  0  0  0  7  0  0  0  0  0  0  0  8  0  0  0  0  0  0  0  9  0  0  0  0  0  0  0  10  0  0  3.6  1.2  0.8  0.2  0  11  0  0  0  0  0  0  0  12  0  0  0  0  0  0  0  13  0  0  4  0  0  0  0  14  0.2  0.4  0  0  0  0  0  15  1  2  0  0  9.4  0.2  0  16  0  0  0  0  0  0  0  17  0  0  0  0  0  0  0  18  0  0  0  0  0  0  0  19  0  0  0  0  0  0  0  20  0  0  0  0  0  0  0  21  0  1  177.6  1.2  2  2.4  1  22  0  0  0  0  0  0  0  23  0  0  0  0  0  0  0  24  0.2  0.4  2.2  1  0.4  0  0  25  0  0.2  0.2  1.6  1  0.8  4.6  26  0  0  0  0  0  0  0  27  0  0  0  0  0  0  0  28  0  0  0  0  0  0  0  29  0  0  0  0  0  0  0  30  0  0.2  2  1.4  0.2  1.2  2.2  31  0  0  0  0  0  0  0      203     No.  Taxon N P1 .0 20 8  N P2 .0 20 8  N P3 .0 20 8  N P4 .0 20 8  N P5 .0 20 8  N P6 .0 20 8  N P7 .0 20 8  N P8 .0 20 8  32  Hydrobiosis frater  1  0  0  0  0  0  0  0  33  Hydrobiosis spatulata  0  0  1.8  0  0  0  0  0  34  Hydrobiosis clavigera  0  0  1  0  0  0  0  0  35  Hydrobiosis harpidosa  0  0  0  0  0  0.4  0  0  36  Hydrobiosis charadrea 0  0  0  0  0  0  0  0  37  Neurochorema forsteri 0  0  0  0  0  0  5  1  38  Neurochorema confusum 0  0  0.2  0  0  0  0  0  39  Costachorema callistum 0.6  0.4  0.6  1.6  0  0.4  2.8  0.6  40  Costachorema xanthopterum 0  0.4  0  0  0  0  0  0  41  Hydrobiosella mixta  0  0  0  0  0  0  0  0  42  Hydrochorema crassicaudatum 0  0  0  0  0.2  0  0  0  43  Caddis Pupae  0  0  0  0  0  0  0.8  0  44  Plectrocnemia machlachlani 0  0  0  0  0  0  0  0  45  Polyplectropus sp.  0  0  0  0  0  0  0  0  46  Pycnocentria evecta  0  0  4  0  0  0  0.2  1.4  47  Pycnocentria funerea  0  0  1  0  0  0  0  1.4  48  Pycnocentria hawdonia 0  0  0  0  0  0  0.8  0  49  Pycnocentria sylvestris 0  0  0  0  0  0  0  0  50  Pycnocentria gunni  0  0  0  0  18.6  0  0.4  0  51  Pycnocentrodes aeris  0  0  0.6  0  1.2  0  1.2  17.4  52  Beraeoptera roria  0  0.6  6.6  5.4  0  0.2  1.6  225.2  53  Confluens hamiltoni  0  0  15.6  1.6  0  0  2.4  1.6  54  Oeconesus similis  0  0  0  0  0  0  0  0  55  Hudsonema amabile  0  0  0.2  0  0  0  0  0  56  Hudsonema alienum  0  0  0  0  0  0  0  0  57  Triplectides dolichos  0  0  0  0  0  0  0  0.2  58  Triplectides obsoletus  0  0  0  0  0  0  0  0  59  Oxyethira albiceps  0.2  0.4  6.8  1.6  0  0  0  0.6  60  Paroxyethira sp.  0  0  0  0  0  0  0  0  61  Olinga feredayi  0  0  1.8  0.4  0  0  0  4.8  62  Helicopsyche poutini  0  0  0.6  0  0  0  0  0.6  63  Pseudoeconesus sp.  0  0  0  0  0  0  0  0  64  Zelolessica cheira  0  0.4  14.6  1.2  1  0.6  2  0.8  65  Allocentrella sp.  0  0  3.4  0  0  0  0  0  Diptera 66  Orthocladiinae A  0.6  0  0.2  0  0  0  0  0.8  67  Orthocladiinae B  0  19.8  0  55.2  12.8  39.2  53.4  8  68  Orthocladiinae C  0  0  4.2  93.4  13.6  4.8  0  0.2  69  Orthocladiinae D  0.2  6.6  0  9.2  0  5  0  0  70  Orthocladiinae E  0  12.2  67  35.8  63.6  2.2  6.4  0.8  71  Orthocladiinae F  0  0  0  0  0  0  0  0  72  Orthocladiinae Z  0  0  0  0  0  0  0  0  204     No.  H B1 .0 20 8  H B2 .0 20 8  H B3 .0 20 8  H B4 .0 20 8  H B5 .0 20 8  H B6 .0 20 8  H B7 .0 20 8  H B8 .0 20 8  N P1 .1 00 8  N P2 .1 00 8  N P3 .1 00 8  N P4 .1 00 8  32  0  0  0  0  0  0  0  0  0  0  0  0  33  0  0  0  0  0  0  0  0  0  0  0  0  34  0  0  0.8  0  0  0  0  0  0  0  0  0  35  0  0  0  0  0  0  0  0  0  0  0  0  36  0  0  0  0  0  0  0  0  0  0  0  0  37  0  0  3.2  0  0  0  0  0  0  0  1.2  0  38  0  0  0  0.4  0  0  0  0  0  0  0  0  39  0.8  0  2.4  0  0  0  0  0  0.4  0.6  0  2.4  40  0  0  0  0  0  0  0  0  0  0  0  0  41  0  0  0  0  0  0  0  0  0  0  0  0  42  0  0  0  0  0  0  0  0  0  0  0  0  43  0  0  0  0  0  0  0  0  0  0  0  0  44  0  0  0  0  0  0  0  0  0  0  0  0  45  0  0  0  0  0  0  0  0  0  0  0  0  46  0.8  4  306.4  6  2.4  12  0.2  63.2  12.64  0  25.6  0  47  0  0  0  0  2.4  0  0  0  0  0.8  2.4  0  48  0  0  0  0  0  0  1  0  0  0  0  0  49  0  0  0  0  0  0  0  0  0  0  0  0  50  0  0  0  0  0  0  0  0  0  0  0  0  51  4.8  6.4  367.2  13.6  88  255.2  20  52.8  10.56  0  0  4  52  0  0  0  0  0  0  0  0  0  12  7.6  72.4  53  0  0  0  0  0  0  0  0  0  0  4.8  0  54  0  0  0  0  0  0  0  0  0  0  0  0  55  0  0  0  0.8  0  0  0.2  2.4  0.48  0  0  0  56  0  0  0  0  0  0  0  0  0  0  0  0  57  0  0  0  0  0  0  0  0  0  0  3.2  0  58  0  0  0  0  0  0  0  0  0  0  0  0  59  0  0  0.8  17.6  0  0.2  37.8  60.8  12.36  0  0  0  60  0  0  0  0  0  0  0  0  0  0  0  0  61  10  11.2  19.2  0.8  8.2  20  0.2  0  0  0  5.6  0  62  0  0  5.6  0  0  0  0  0  0  0.2  0  0  63  0  0  0  0  0  0  0  0  0  0  0  0  64  0  0  0  0  0  0  0  0  0  0.4  26.8  0  65  0  0  0  0  0  0  0  0  0  0  6.4  0  66  0  0  0  0  0.8  0  0  0  0  0  2.4  1.6  67  0  0  0  0  0  0  0  0  0.4  0.2  21.2  34  68  0  2.4  22.4  5.6  2.4  13.6  5.2  0.8  0.16  4.2  0  10.4  69  0  0  0  0  0  0  0  0  0.4  0.8  2.8  110.4  70  22.8  13.6  12.8  152.4  12.8  10.4  19.2  7.2  1.44  0.6  33.2  24.8  71  0  0  0  11.2  0  0  0  2.4  0.48  0  0  0  72  0.8  0  7.2  0  0  0  0.8  0  0  0  0  0      205     No.  N P5 .1 00 8  N P6 .1 00 8  N P7 .1 00 8  N P8 .1 00 8  H B1 .1 00 8  H B2 .1 00 8  H B3 .1 00 8  H B4 .1 00 8  H B5 .1 00 8  H B6 .1 00 8  H B7 .1 00 8  H B8 .1 00 8  32  0  0.4  0  0  0  0  0  0  0  0  0  0  33  0  0  0  0  0  0  0  0  0  0  0  0  34  0  0  0  0  0  0  0  0  0  0  0  0  35  0  0  0  0  0  0  0  0  0  0  0  0  36  0  0  0  0  0  0  0  0  0  0  0  0  37  0  0  0.2  0  0  0  0.2  0  0  0  0  0  38  0  0  0  0  0  0  0.2  0.2  0  0  0  0  39  0  1  0  0.2  0.6  0  2.6  0  0  0  0  0  40  0  0  0  0  0  0  0  0  0  0  0  0  41  0  0  0  0  0  0  0  0  0  0  0  0  42  0  0  0  0  0  0  0  0  0  0  0  0  43  0  0  0  0  0  0  0  0  0  0  0  0  44  0  0  0  0  0.2  0.2  0.2  0  0  0  0  0  45  0  0  0  0.2  0  0  0  0  0  0.2  0  0.4  46  0  0  0  0  0  0  19  0.2  0.2  0.4  27.2  8.2  47  0.2  0  0  1.2  0  0  0  0  0  0  0  0  48  0  0  0  0  0  0  0  0  0  0  0  0  49  0  0  0  0  0  0  0  0  0  0  0  0  50  0  0  0  0  0  0  0  0  0  0  0  0  51  0  0  0.8  0.2  18.6  6.4  120.8  3  28.2  106  432  36.2  52  0  0  24.8  23  0.8  0.8  0  0  0  0  0  0  53  0  0  0.2  0.2  0  0  0  0  0  0  0  0  54  0  0  0  0  0  0  0  0  0  0  0  0  55  0  0  0  0  0  0.2  0  0  0  0.4  11  0  56  0  0  0  0  0  0  0  0  0  0  0  0  57  0  0  0  0  0  0  0  0  0  0  0  0  58  0  0  0  0  0  0  0  0  0  0  0  0  59  0.2  0  0  0  0.2  0.2  0  2  0  0  0  0.2  60  0  0  0  0  0  0  0  0  0  0  0  0  61  0  0  0.4  0  1.2  0.6  9.8  0  0.4  0.6  0  0  62  0  0  0  0  0  0  19  0  0  0  0  0  63  0  0  0  0  0  0  0  0  0  0  0  0  64  0  0  0.2  0  0  0  0  0  0  0  0  0  65  0  0  0  0  0  0  0  0  0  0  0  0  66  0  8.2  2.2  0  0  0  0  0  0  0  0  0  67  19.6  18.2  18.8  118.6  0  1.2  71.2  163  0  92  9.8  5.4  68  0  2.4  0  0  0  0  0  0  0  0  0  0  69  0.8  0.4  39.2  12.6  1.2  0.8  16.4  233.8  0  61.4  3.2  1.6  70  0.2  4.8  6.2  19  0.4  0.6  3.2  181.6  0  0.6  8  5.2  71  0  0  0  0  0  0  0  0  0  0  0  0  72  0  0  0  0  0  0.2  0  0.6  0  0  0  0      206     No.  N P1 .0 20 9  N P2 .0 20 9  N P3 .0 20 9  N P4 .0 20 9  N P5 .0 20 9  N P6 .0 20 9  N P7 .0 20 9  N P8 .0 20 9  H B1 .0 20 9  H B2 .0 20 9  H B3 .0 20 9  H B4 .0 20 9  32  0  0  0  0.2  0  0  0  0  0.4  0  0  0  33  0  0  2  0  0  0  0  0.2  0  0  0  0  34  0  0  0  0  0  0  0  0  0  0  2.4  0.2  35  0  0  0  0  0  0  0  0  0  0  0  0  36  0.4  0  0  0  0  0  0  0  0  0  0  0  37  0  0.8  0  0  0  0  0.8  0.2  0  0  24  0  38  0  0  0  0  0  0  0  0  0  0  0  0  39  0  0  0.8  0  0  0  1  5.4  0  0  4  0  40  0  0  0  0  0  0  0  0  0  0.2  2.4  0  41  0  0  0  0  0  0  0  0  0  0  0  0  42  0  0  0  0  0  0  0  0  0  0  0  0  43  0  0  0  0  0  0  0  0  0  0  0  0  44  0  0  0  0  0  0  0  0  0  0  0  0  45  0  0  0  0  0  0  0  0  0.2  0  0  0  46  0  0  0  0  0  0  0  0  3.4  0.4  300  15.4  47  0  0  0  0  0  0  0.4  4.8  0  0  0  0  48  0  0  0.6  0  0  0  0  0  0  0  0  0  49  0  0  5.8  0  0  0  0  0  0  0  0  0  50  0  0  0  0  0  0  0  0  0  0  0  0  51  0.2  0  2.8  1.4  0.8  0  1  0.2  43  11.2  380.8  34.4  52  0  2.8  1.6  1.8  0  0.2  1.4  50  1.6  0  0  0  53  0  0  13  0  0  0  0  0  0  0  0  0  54  0  0.2  0  0  0  0  0  0  0  0  0  0  55  0  0  0  0  0  0  0  0.2  0  0  0  0  56  0  0  0  0  0  0  0  0  0  0  0  0  57  0  0  0  0  0  0  0  0  0  0  0  0  58  0  0  0  0  0  0  0  0  0  0  0  0  59  0  1.6  0  0  0.2  0  0  0.8  0.8  0  0  1.6  60  0  0  0  0  0  0  0  0  0  0  0  0.8  61  0  1  0.2  0  0  0  0  4.4  12.4  4  66.4  0  62  0  0.2  0  0  0  0  0  0  0  0  20.8  0  63  0  0  0  0  0  0  0  0  0  0  0  0  64  0  1.2  12  0.6  0  0  0  1  0  0  0  0  65  0  0  4.2  0  0  0  0  0  0  0  0  0  66  0  0  0  0  0  0  0.8  0  0  0  0  0  67  4  40.8  12  2  115  2.4  55.8  25.8  17  35.6  47.2  148.4  68  0  4.6  0  0  0  0  0  0  0  0  0  0  69  0  2.4  1.8  0  45.8  0.2  2  0.2  0  5  4  9.4  70  0.4  2.4  3  0  32.6  1.2  0.6  2.8  0  1.8  0  9.2  71  0  0  0  0  0  0  0  6  0  0  0  12.2  72  0  0  0  0  0  0  0  0  0  0  0.8  0      207     No.  H B5 .0 20 9  H B6 .0 20 9  H B7 .0 20 9  H B8 .0 20 9  N P1 .0 70 9  N P2 .0 70 9  N P3 .0 70 9  N P4 .0 70 9  N P5 .0 70 9  N P6 .0 70 9  N P7 .0 70 9  N P8 .0 70 9  32  0  0  0  0  0  0  0  0  0  0  0  0  33  0  0  0  0  0  0.2  0.8  0  8.2  0  0  0.4  34  0.2  0  0  0  0  0  0  0  0  0  0  0  35  0  0  0  0  0  0  0  0  0  0  0  0  36  0  0  0  0  1  0  0  0  0  0  0  0  37  0  0  0  0  0  0.4  0  0  0  0  3.8  2  38  0  0  0  0  0  0  0  0  0  0  0  0  39  0  0  0  0  0.2  0  0.4  1.6  0.2  1  0.8  2  40  0  0  0  0  0  0  0  0  0  0  0  0.4  41  0  0  0  0  0  0  0.2  0  0  0  0  0  42  0  0  0  0  0  0  0  0  0  0  0  0  43  0  0  0  0  0  0  0  0  0  0  0  0  44  0.2  0  0  0  0  0  0  0  0  0  0  0  45  0  0  0  1.2  0  0  0  0  0  0  0  0  46  8.4  2.8  0  12.4  0  0  0  0  0  0  0  0  47  0  0  0  0  0  0  0.8  0.8  2.2  0  6  5.4  48  0  0  0  0  0  0  4.6  0  0  0  0  0  49  0  0  0  0  0  0  0  0  0  0  0  0  50  0  0  0  0  0  0  0  0  0  0  0  0  51  95.8  92.4  1.6  14.2  0  0  0.6  0.2  1.2  0  6.4  3.4  52  0  0  0  0  0  10.2  6.4  1.8  0  1  110.6  138.8  53  0  0  0  0  0  0  1.2  0.2  0  0  0  0  54  0  0  0  0  0  0  0  0  0  0  0  0  55  0.4  0.2  0.8  0  0  0  0  0  0  0  0  0.2  56  0  0  0  0  0  0  0.4  0  0  0  0  0  57  0  0  0  0  0  0  0  0  0  0  0  0.2  58  0  0  0  0.2  0  0  0  0  0  0  0  0  59  0  0  2  4.8  0  0.2  0  0.8  0.2  0  0  0.2  60  0  0  0  0  0  0  0  0  0  0  0  0.2  61  8.2  6.4  0  0  0  0  0.2  0  0  0  0  0.8  62  0  0  0  0  0  0  0  0  0  0  0  0  63  0  0  0  0  0  0  0  0  0  0  0.2  0  64  0  0  0  0  0  0.8  12.6  0.8  0  0  0  0  65  0  0  0  0  0  0  2.4  0  0  0  0  0  66  0  0  0  0  0.2  0  0  0  0  11.2  5.4  0  67  25.8  1.6  199.4  17.2  166.6  18.8  6  186.2  32.8  124.8  19.4  1  68  0  0  0  0  0  28.2  0  0  9  0  0  0  69  5.4  314.4  43.8  19.2  0.8  0  0  8  2.6  0  5.2  0  70  0  0  12.4  9  0  7.8  0  20.2  32.6  0  24  2.4  71  0  0  9.2  0  0  0  0  0  0  0  0  1  72  0  0  0  20.4  0  0  0  0  0  0  0  0      208     H B1 .0 70 9  No.  H B2 .0 70 9  H B3 .0 70 9  H B4 .0 70 9  H B5 .0 70 9  H B6 .0 70 9  H B7 .0 70 9  32  0  0.2  0  0  0  0  0  33  0  0  0  0  0  0  0  34  0  0  0  0  0  0  0  35  0  0  0  0  0  0  0  36  0  0  0  0  0  0  0  37  0  0  13  0  0  0  0.2  38  0  0  0  0  0  0.2  0  39  0  0  0  0  0  0  0  40  0  0  1.4  0  0  0  0  41  0  0  0  0  0  0  0  42  0  0  0  0  0  0  0  43  0  0  0  0  0  0  0  44  0  0  0  0  0  0  0  45  0  0  0  0.6  0  0  0  46  0  0  146.6  9  1  1  17.2  47  0  0  0  0  0  0  0  48  0  0  0  0  0  0  0  49  0  0  0  0  0  0  0  50  0  0  0  0  0  0  0  51  1.4  1.2  338.2  44.4  0.6  5.4  196.6  52  0.2  0  0  0  0  0  0  53  0  0  0  0  0  0  0  54  0  0  0  0  0  0  0  55  0  0  0  1  0.4  0  16.4  56  0  0  0  0  0  0  0  57  0  0  0  0  0  0  0  58  0  0  0  0  0  0  0  59  0  0  0  0  0  0  0  60  0  0  0  0  0  0  0  61  0  0.4  13.8  0  1.6  1.2  1.4  62  0  0  16  0  0  0  0  63  0  0  0  0  0  0  0  64  0  0  0  0  0  0  0  65  0  0  0  0  0  0  0  66  0  0  0  0  0  0.2  0  67  1.6  0.4  44.4  7.2  2.8  1  0.8  68  0  0  0  0  0  0  0  69  0.4  0  7  1.8  0  0.6  0  70  0  0  0  3.4  0  0  0  71  0  0  0.8  0  0  0  0.2  72  0  0  0  0  0  0  0      209     No.  Taxon N P1 .0 20 8  N P2 .0 20 8  N P3 .0 20 8  N P4 .0 20 8  N P5 .0 20 8  N P6 .0 20 8  N P7 .0 20 8  N P8 .0 20 8  73  Chironominae A  0  11.8  2  4.6  16  1.8  9  1.6  74  Chironominae B  0  0  0  0  0  0  0  0.8  75  Maoridiamesa sp.  0  101.6  1  11.6  176.2  17.2  92.6  8.6  76  Tanypodinae  0  0  1.4  0.8  0.8  0  0  0.4  77  Chironomidae pupae  0  4.2  3.4  2.8  1.2  0.4  2.6  0.2  78  Eriopterini (other)  0  5.8  0  1.6  0.4  3.2  4.2  0.6  79  Molophilus sp.  1.2  0  0  0  0  0  0  0  80  Aphrophila neozelandica 0  0.6  3.6  7.2  0.2  0.8  1.2  13.8  81  Austrosimulium spp.  0  15.6  1.8  0  3  1.8  0.2  0  82  Mischoderus sp.  0  0.2  0  0  0  0  0.4  2.4  83  Empididae  0  2.6  0  0  7.4  3.6  0  0  84  Ceratopogonidae  0  0  0  0  0  0  0  0  85  Limonia sp.  0  0  0  0  0.2  0  0  0  86  Ephydridae  0  0  0  0  0  0  0  0.2  87  Muscidae  0  0  0.2  4.8  2.4  3.4  5.8  0  88  Paralimnophila sp.  0  0  0  0  0  0  0  0  89  Stratiomyidae  0  0  0.4  0  0  0  0  0  90  Unidentified Diptera  0  0  0  0  0  0  0  0  Coleoptera  91  Elmidae Hydora spp.  0  6.6  15.2  42  0.4  1.6  35.6  15.4  92  Hydraenidae  0  0  6  0.2  0  0  0  1.6  93  Hydrophylidae  0  0  0.2  0  0  0  0  0  94  Ptilodactylidae  0  0  5  0.2  0  0  0  0  95  Scirtidae  0  0  0  0  0.2  0  0  0  Mollusca  96  Potamopyrgus antipodarum 0  0  14.4  0  0  0  0  0.2  97  Physa spp.  0  0  0  0  0  0  0  0.2  98  Ferissia spp.  0  0  0  0  0  0  0  0  99  Latia neritoides  0  0  0  0  0  0  0  0  100  Gyraulus corinna  0  0  0  0  0  0  0  0  101  Sphaerium novaezelandiae 0  0  0  0  0  0  0  0  Other  102  Archichaulioides diversus 0  0  15.6  0.4  0  0  0.2  64.2  103  Platyhelminthes  0  1.4  0.8  0  0.2  0.2  0  2.8  104  Oligochaeta  0  0.2  2  4.8  59.4  55.4  0.4  1.4  105  Nematomorpha  0  0  0  0  0.2  0  0  0  106  Paricalliope spp.  0.2  0  0  0  0  0  0  0  107  Ostracoda  0  0  0  0  0  0  0  0  108  Hirudinea   0  0  0  0  0  0  0  0  109  Nemertea  0  0  0  0  0  0  0  0  110  Nematoda  0  0  0  0  0  0  0  0  111  Xanthocnemia zelandica 0  0  0  0  0  0  0  0      210     No.  H B1 .0 20 8  H B2 .0 20 8  H B3 .0 20 8  H B4 .0 20 8  H B5 .0 20 8  H B6 .0 20 8  H B7 .0 20 8  H B8 .0 20 8  N P1 .1 00 8  N P2 .1 00 8  N P3 .1 00 8  N P4 .1 00 8  73  1.6  1.6  15.2  176.6  7.2  61.6  27.8  24  4.8  5.2  4.8  8  74  0  0  0  0  0  0  0  0.2  0.04  0  0  0  75  0  0  0.2  0  0  0  3.2  0  0  0.8  0  47.2  76  0  0  0  0  0.8  0  0  0  0  0  0.8  0  77  0  0  4  32.8  0  5.6  1.8  1.6  0.32  0  2.4  3.2  78  0.4  2.4  0  0.8  6.6  12  0  0  0  0.8  0  5.2  79  0  0.2  0  0  0.4  0  0  0  0  0  0  0  80  0  0  0  0  1.4  3  0  0  0  0.4  9.2  4.6  81  17.6  8  0  0  2.4  0  2.6  1  0.2  0  0  0  82  0  0  0  0  0  0  0  0  0  0  0  0  83  0  0  0  0  0  0  0  0  0  0  2  0  84  0  0  0  0  0  0  0  0  0  0  0.4  0  85  0  0  0  0  0  0  0  0  0  0  0  0  86  0  0  0  0  0  0  0  0  0  0  0  0  87  0  0  0  0  0  0  0  0  0  0  2.4  0.2  88  0  0  0  0  0  0  0  0  0  0  0  0  89  0  0  0  0  0  0  0  0  0  0  0  0  90  0  0  0  0  0  0  0  0  0  0  0  0  91  37.8  20.8  1.6  25.6  55.2  32.8  0.2  0  0  1.2  25.2  46  92  0  0  0  0  0  0  0  0  0  0  4  0  93  0  0  0  0  0  0  0  0  0  0  0  0  94  0  0  0  0  0  0  0  0  0  0  6.4  0  95  0  0  0  0  0  0  0  0  0  0  0  0  96  0.8  0  7.2  43.2  8.8  5.4  1175  40.8  8.16  0  4  0  97  0  0  0  3.2  0  0.2  70.8  2  0.4  0  0  0  98  0  0  0  64  0  0  0  7.2  1.44  0  0  0  99  0  0  0  0  0  0  0  0  0  0  0  0  100  0  0  0  0.2  0  0  2.4  0  0  0  0  0  101  0  0  0  0  0  0  0  0  0  0  0  0  102  0.2  0.8  1.6  1.6  3.4  1.6  0  0  0  0  21.6  0  103  0  0  19.2  16  0.8  8  25.4  4.2  0.84  0  0  0  104  1.6  5.6  12  44.8  9.6  7.2  692  16.8  3.36  0  0.8  0  105  0  0  0  0  0  0  0  0  0  0  0  0  106  0  0  2.6  66.4  0  0.8  15.2  53.6  10.72  0  0  0  107  0.4  0.8  0.8  0.8  0  1  2.6  0  0  0  0  0  108  0  0  0  0.8  0  0  0.4  0  0  0  0  0  109  0  0  0  0  0  0  0  2.4  0.48  0  0  0  110  0  0  0  0  0  0  0  0  0  0  0  0  111  0  0  0  0  0  0  0  0  0  0  0  0      211     No.  N P5 .1 00 8  N P6 .1 00 8  N P7 .1 00 8  N P8 .1 00 8  H B1 .1 00 8  H B2 .1 00 8  H B3 .1 00 8  H B4 .1 00 8  H B5 .1 00 8  H B6 .1 00 8  H B7 .1 00 8  H B8 .1 00 8  73  2  3.2  19.8  6.8  0.4  0.4  2.4  1.2  0.4  7.4  0  0  74  0  0  0  0  0  0.4  0  5.6  0  1.4  0  0  75  15.4  5  14.4  20  0  0  9.6  1  0  0  0  0.2  76  0  0  0  0  0  0.4  0  4.6  0.2  3  0  0  77  1.8  1.8  10  15  0.2  0.2  2.6  70.2  0.6  10.8  0.8  0  78  0.8  2.6  3.2  1  1  1  0  0  2  3.4  0  0  79  0  0.8  0  0  0  0  0  0  0  0  0  0  80  0  3.4  5.8  2.2  0.6  1.8  0  0  3.4  7.2  0  0  81  0.2  0  0  0  1.2  3  0  0  0.6  5.6  4  0.2  82  0  0  0  0  0  0  0  0  0  0  0  0  83  0  0  0  0  0  0  0  0  0  0  0  0  84  0  0  0  0  0  0  0  0  0  0  0  0  85  0  0  0  0  0  0  0  0  0  0  0  0  86  0  0  0  0  0  0  0  0  0  0  0  0  87  0.8  0.6  0  0  0  0  0  0  0  0  0  0  88  0  0  0  0  0  0  0  0  0  0  0  0  89  0  0  0  0  0  0  0  0  0  0  0  0  90  0  0  0  0  0  0  0  0  0  0  0.2  0  91  0  0.4  15.8  3  3  2.6  0  0.4  2.4  7.6  0  0  92  0  0  0  0  0  0.2  0  0  0  0  0  0  93  0  0  0  0  0  0  0  0  0  0  0  0  94  0  0  0  0  0  0  0  0  0  0  0  0  95  0  0  0  0  0  0  0  0  0  0  0  0  96  0  0  0  0  0  0  7.4  16  0.2  0  135.8  2  97  0  0  0  0  0  0  0  0.2  0  0  4  0  98  0  0  0  0  0  0  0  0  0  0  0  0  99  0  0  0  0  0  0  0  0  0  0  0  0  100  0  0  0  0  0  0  0  0  0  0  0  0  101  0  0  0  0  0  0  0.8  0  0  0  0  0  102  0  0  0  0  0  0.2  1.8  0  0  0.2  0  0  103  0  0.2  0  0  0  0  36.2  0.4  0  0  1.6  0  104  0.8  0  0  0.4  0  0.2  5.6  108.8  0  32.6  83.4  17  105  0  0  0  0  0  0  0  0  0  0  0  0  106  0  0  0  0  0  0  6.4  81.6  0  0  63.4  552.6  107  0  0  0  0  0  0  0  0  0  0  0  0.2  108  0  0  0  0  0  0  0  0  0  0  0  0  109  0  0  0  0  0  0  0  0  0  0  0  0  110  0  0.2  0  0  0  0  0  0  0  0  0  0  111  0  0  0  0  0  0  0  0.2  0  0  0  0      212     No.  N P1 .0 20 9  N P2 .0 20 9  N P3 .0 20 9  N P4 .0 20 9  N P5 .0 20 9  N P6 .0 20 9  N P7 .0 20 9  N P8 .0 20 9  H B1 .0 20 9  H B2 .0 20 9  H B3 .0 20 9  H B4 .0 20 9  73  0.2  2.8  0.8  0  12.2  0.4  2  3.8  1.4  5.8  4.8  43  74  0  0  0  0  0  0  0  0  0  0  0  0  75  0.4  66.8  0.2  0.4  328.6  2  11.8  1.2  0  0  8  1.2  76  0  0  0.4  0  0  0  0  0  0  0  0  0  77  0  0.6  0  0  12.2  0  3.2  1.8  0.4  0.4  2.4  13.8  78  0  5.4  0  1  0.2  0  6.6  1.2  2.8  2.2  0  1  79  0  0  0  0  0  0  0  0  0  1.2  0  0  80  0  0  4  0.6  0  0  5.8  13  0.2  0.2  0  0  81  0  14.2  0.4  0  2.8  0  0.2  1.8  11.2  1  0  0  82  0  0.2  0  0  0  0  0  0.2  0  0  0  0  83  0  0  0  0  2  0.2  1.4  0  0  0.4  0  0  84  0  0  0  0  0  0  0  0  0  0  0  0  85  0  0  0  0  0  0  0  0  0  0  0  0  86  0  0  0  0  0  0  0  0  0  0  0  0  87  0  0.6  0  0  5  0  0.4  0  0  0  0  0  88  0  0  0  0  0  0  0  0  0  0  0  0  89  0  0  0  0  0  0  0  0  0  0  0  0  90  0  0  0  0  0  0  0  0  0  0  0  0  91  0  9.2  10.8  18  0  0  20.4  2.4  6.4  11.4  0  13.4  92  0  0  4.6  0  0  0  0  2.8  0  0  0  0  93  0  0  0  0  0  0  0  0  0  0  0  0  94  0  0  0.4  0  0  0  0  0  0  0  0  0  95  0  0  0  0  0  0  0  0  0  0  0  0  96  0  0  2.8  0  0  0  0.8  0  0  0.6  34.4  171.4  97  0  0  0  0  0  0  0  0  0  0  0  6.4  98  0  0  0  0  0  0  0  0  0  0  0  0  99  0  0  0  0  0  0  0  0  0  0  0  32.8  100  0  0  0  0  0  0  0  0  0  0  0  0  101  0  0  0  0  0  0  0  0  0  0  0  0  102  0  0.2  9.2  0.2  0  0  1.2  5.2  3  0.4  0.4  3.2  103  0  1.2  0  0  0.8  0  0  0.8  0  0  12  10.6  104  0  0.2  4.6  0.6  23.4  0.2  4.8  1.8  4.6  4.8  12.8  6.8  105  0  0  0  0  0  0  0  0  0  0  0  0  106  0.6  0  0  0  0  0.2  0  0  0.2  0.4  5.6  5.6  107  0  0  0  0  0  0  0  0  0  0  0  0  108  0  0  0  0  0  0  0  0  0  0  0  0  109  0  0  0  0  0  0  0  0  0  0  0  0.8  110  0  0  0  0  2.4  0  0  0  0  0  0  0  111  0  0  0  0  0  0  0  0  0  0  0  0      213     No.  H B5 .0 20 9  H B6 .0 20 9  H B7 .0 20 9  H B8 .0 20 9  N P1 .0 70 9  N P2 .0 70 9  N P3 .0 70 9  N P4 .0 70 9  N P5 .0 70 9  N P6 .0 70 9  N P7 .0 70 9  N P8 .0 70 9  73  3  444.8  106.2  18  0  0  0  0  6.6  0  3.2  0.4  74  0  0  0  9.2  0  0  0  0  0  0  0  0  75  0.8  0  0.8  0  24.8  26  1  88.6  168.6  16.8  63.4  19.2  76  0  0  0  1.6  0  0  0  0  0  0  0  0  77  1  32.4  16.8  1.4  2.6  0  0  2.4  0.8  3  0.2  0.4  78  2.2  5.8  0  0  0  1.4  0  2.2  0.6  2.8  4.8  2.8  79  0  0  0  0  1  0  0  0  0  0  0  0  80  1.2  2  0  0  0  0.4  1  0.8  0  0  5  5.4  81  1.6  1  1.2  1.6  0.2  12.4  0  0  1  7  0.8  0  82  0  0  0  0  0.2  0  0  0  0  0  0.2  0  83  0  0  0  0  0  0  0.2  0  0  0  0.8  0  84  0  0  0  0  0  0  0  0  0  0  0  0  85  0  0  0  0  0  0  0  0  0.2  0  0  0  86  0  0  0  0  0  0  0  0  0  0  0  0  87  0  0.6  0  0  0  0.8  0  0.4  4.2  0  0  0.2  88  0  0  0  0  0  0  0  0  0  0  0  0.2  89  0  0  0  0  0  0  0  0  0  0  0  0  90  0  0  0  0  0  0  0  0  0  0  0  0  91  42  47.8  0  0  0  0.6  4.8  1.6  0  1  19.4  1  92  0  0  0  0  0  0  0.6  0  0  0  0  1  93  0  0.2  0  0  0  0  0.6  0  0  0  0  0  94  0  0  0  0  0  0  0.4  0  0  0  0  0  95  0  0  0  0  0  0  0  0  0  0  0  0  96  11.6  17.4  453  52.8  0  0  0  0  0  0  0  0  97  0  0  10.8  0  0  0  0  0  0.2  0  0  0  98  0  0  0  0  0  0  0  0  0  0  0  0  99  0  0  0  14.4  0  0  0  0  0  0  0  0  100  0  0  0  0  0  0  0  0  0  0  0  0  101  0  0  0  1  0  0  0  0  0  0  0  0  102  1.8  3.8  0  0  0  0.2  0.4  0  0  0  0  1  103  1  16.8  2.4  5.6  0  0  0  0  0  0  0.2  0.2  104  12  3.4  524.6  132.6  0.2  0.2  0.4  2.6  0.8  0  12  0.2  105  0  0  0  0  0  0  0  0  0  0  0  0  106  0  0  4.8  121.2  0  0  0  0  0  0  0  0  107  0  6.6  0  0.4  0  0  0  0  0  0  0  0  108  0  0  0.2  0  0  0  0  0  0  0  0  0  109  0  2  0  2  0  0  0  0  0  0  0  0  110  0  0  0  0  0  0  0  0  0  0  0  0  111  0  0  0  0  0  0  0  0  0  0  0  0      214     H B1 .0 70 9  No.  H B2 .0 70 9  H B3 .0 70 9  H B4 .0 70 9  H B5 .0 70 9  H B6 .0 70 9  H B7 .0 70 9  73  0  0  8  0.2  0  0  0.2  74  0  0  0  0  0  0  0  75  2.2  0.2  12.2  0.6  0  0  0  76  0  0  0  1.8  0  0  0  77  0  0  0.8  0.4  0  0  0  78  0.2  0  0  0.2  1  1  0  79  0.2  0.2  0  0  0.2  0  0  80  0  0.2  0  0  0  0.2  0  81  0  0.4  0  0.4  0.2  0.8  0.2  82  0  0  0  0  0  0  0  83  0  0  0  0  0  0  0  84  0  0  0  0  0  0  0  85  0  0  0  0  0  0  0  86  0  0  0  0  0  0  0  87  0  0  0  0  0  0  0  88  0  0  0  0  0  0  0.2  89  0  0  0  0  0  0  0  90  0  0  0  0  0  0  0  91  0  0  0.2  0.4  0.4  0  0  92  0  0  0  0  0  0  0  93  0  0  0  0  0  0  0  94  0  0  0  0  0  0  0  95  0  0  0  0  0  0  0  96  0  0  2  39.2  0  0.2  102  97  0  0  0  1.8  0  0  4  98  0  0  0  0  0  0  0  99  0  0  0  0.2  0  0  0  100  0  0  0  0  0  0  0  101  0  0  0  0  0  0  0  102  0.2  0.4  0  0  0.4  0  0  103  0  0  27  7  0  0  1  104  0  0.6  4.6  4.4  0  0  33.8  105  0  0  0  0  0  0  0  106  0  0.6  9  54.2  0  0  60.2  107  0  0  0.2  0  0  0  0  108  0  0  0  0  0  0  0.4  109  0  0  0  0  0  0  0  110  0  0  0  0  0  0  0  111  0  0  0  0  0  0  0      215     Appendix 4: Raw  invertebrate data collected  from 24 streams of Cantabria, Spain,  July 2007  (Chapter 4).  Values are mean number of individuals collected from five Surber samples (area =  0.1 m2) at each site.  E = Ephemeroptera, P = Plecoptera, TC = cased caddisfly, TF = free‐living  caddisfly, TN = net‐spinning caddisfly, D = Diptera, DC = Chironomidae, CA = adult Coleoptera,  CL = larval Coleoptera, O = Other.  The taxa contributing the most to group differences are as  follows:  D1  =  Prosimulium  spp.,  E1  =  Baetis  spp.,  O1  =  Echinogammarus  spp.,  DC4  =  Orthocladiinae  b,  E3  =  Ephemerella  sp., O3  =  Bythinella  sp.,  P1  =  Protonemura  sp.,  TN1  =  Hydropsyche sp., DC1 = Orthocladiinae a, O2 = Oligochaeta.  Site names are given in Appendix  1 of Chapter 4.    Taxon 1 2  3  4 5 6 7 8 9  10  11 12 Ephemeroptera E1  280.4  335.2  150.8  73  189.8  1492  73.6  65.8  128  81.8  158.4  238.2  E2  3.2  1.8  4.8  0.2  19  8  2.4  13.4  12.8  9.6  1  3.8  E3  18  49.2  80.4  39.2  36  19.2  2.2  0  4.8  1  1  6.2  E4  0  0  0  0  0  0  0  0  0  0  0  0  E5  0  0  0  0  0  0  0  0  0  0  0  0  E6  0  0  0  0  1.6  0  1.2  0.4  0.4  0.4  0  0.4  E7  0  0  0  0  0  0  0.4  0  0  0  0  0  E8  0  0  0  0  0  0  0  0  0  0  0  0.8  Plecoptera P1  4.4  8.6  4.8  4.8  1.6  0  6.4  1  28.4  18.8  4  3  P2  0  0  0  0  1  0  0  0  0.4  0  0  0  P3  0  0  0  0.8  0  0  0  0  0  0  0  0  Trichoptera TC1  0  0  0  1.6  0  0  0  0  0  0.2  0.8  0  TC2  0  0  0  0  0  0  0  0  0  0  0  0  TC3  0  0  0  0  0  0  0  0  0  0  0  0  TC4  3.2  0  0.8  0  0.8  0  0  0  0  0  0  0  TC5  0  0  0  0  0  0  0  0  0  0  0  0  TC6  0  0  0  0.2  0  0  0  0  0  0  0  0  TC7  0  0  0  0  0  0  0  0  0  0  0  0  TC8  0  0  0  0  0  0  0  0  0  0  0  0  TC9  0  0.2  0  0  0  0  0  0  0  0  0  0  TC10  0  0.4  0  0  0  0  0  0  0  0  0  0  TF1  12  13.2  6.8  10.2  2.6  3.8  1.4  2  3.6  4.4  4.6  1.2  TF2  0  0  0  0  0  0  0  0  0.8  0  0  0  TF3  0  0.8  0  0  0  0.2  0.8  0  1.8  0  0  0.2  TF4  0.8  0  0  0  0  0  0.8  0  0.8  0  0  0  TF5  0  0  0  0  0  0  0  0  0  0  0  0  TF6  0  0  0  0  0  0  0  0  0  0  0  1  TF7  0  0  0  0  0  0  0  0.2  0  0  0  0.8  TF8  0  0  0  0  0  0  0  0  0  0  0  0  TN1  21.6  16.6  0.8  21.2  1.4  0  12.6  0  3.8  6.6  2.6  0.4  216                       Taxon 13  14  15  16 17 18 19 20 21  22  23 24 Ephemeroptera E1  146  232.2  197.2  200.2  377  433.6  327.2  109.2  706  348.4  93.8  309.4  E2  1  8  4  4.4  5.2  12.8  8.2  15.6  16.2  1.8  13.6  19.6  E3  5.4  3.8  13.8  7.2  73.2  4.8  27.4  11  9  27.4  16.4  28.6  E4  0  0  0.8  0  0  0  0  0  0  0  0  0  E5  4.4  3.2  0  0  0  0  0.2  0  0  0  30  4  E6  0.8  0.2  0  0  0  0  0.4  0  0.8  0  0  0  E7  0.2  0  0  0  0.2  0  0  0  0  0  0  0  E8  1.6  0  2  0  0  0  0.2  0  0  0  0  0  Plecoptera P1  50.6  32.4  24  11.8  19  1.8  34.8  13.2  0.8  42.8  18.2  10.2  P2  0.2  0  0  0  0.2  0  0  1.2  0  1.4  16  2.8  P3  0  0  0  0  0  0  0  0  0  2  0  5.2  Trichoptera TC1  9.6  0.2  0  0  0  0.2  0  0  0  0  0  0  TC2  0  0  0  0  0  0  0  0  1.2  0  0  0.4  TC3  0  0  0  0  0  0  0  0  0.2  2.4  5  0  TC4  0  0  0  0  2.4  0  0  0  0  0.8  1.6  0  TC5  0  0  0  0  0  0  0  0  0  0  2.4  2.4  TC6  0  0  0  0  0  0  0  0  0  0  0  0  TC7  0  0  0  0  0  0  0  0  1.6  0  0  0  TC8  0  0  0  0  0  0  0  0  0.6  0  0  0  TC9  1.6  0  0.2  0  0  0  0  0  0  0  0  0  TC10  0  0  0  0  0  0  0  0  0  0  0  0  TF1  10  3.6  2.4  4.2  10.4  14.2  7  6  7.6  3.2  4.6  6.2  TF2  0  0  0  0  0  0  0  0  0  0  0  0  TF3  6.2  0.6  0.4  1  1.8  0  2.4  2.4  1.2  0  0  0  TF4  0  0.4  0  0  1.6  0  0  0  0  8.4  0  0  TF5  0  0  0  0  0  0  0  0  0  0  0  0.8  TF6  0  0.2  0  0  0  0.2  0  0  0  0  0  0.8  TF7  0.8  1.6  0  0  0  0  0  8.4  0  0  0  0  TF8  0  0  0  0  0  0  0  0  0  0  0  0.2  TN1  27  6.8  14.8  8.8  105.2  13.4  35.8  7.8  0.8  2.2  3.8  2.8  217     Taxon 1 2  3  4 5 6 7 8 9  10  11 12 Diptera  D1  83.2  155  50.8  0.8  50.8  6802  32.4  129.4  17  452.8  100.4  91.4  D2  14  7.4  1.6  0  0  0  0.8  0  0  0.4  2.6  0.4  D3  0  0  0  0  0  6.4  0  0  0.2  0  0  0  D4  0  0.8  0  0  0  0  0.8  0  0  0  0  1.6  D5  0.8  0  0.4  0  0.8  0  0  3.2  0.4  10.4  0  0.2  D7  0  0  0  0  0  0.2  0  0  0  0  0  0  D8  0  0  0.8  0  0.2  0  0  0  0  0  0.2  0  D9  0  0  0  0  0  3.2  0  0  0  0  0  0  D10  0  0.8  0  0.8  0  0  0  0  0  0  0  0  D11  0  0  0  0.2  0  0  0  0  0  0  0  0  D12  0  0  0  0  0  0  0  0  0  0  0  0  DC1  10.4  0.2  9.2  1.4  3.2  423.8  0.8  0.2  1.8  1  30.4  0.2  DC2  4  1.6  0  0  0  28.8  3.2  0  1.4  2.2  1.6  0  DC3  0.4  0.8  2.4  0  3.2  4  6  0.2  1.4  2.4  0  1  DC4  6.8  39.2  0.8  5.4  3.2  1111  32.6  0.6  1.8  17  76.2  4.6  DC5  0  0  0  0  0  0  26.4  0  0  0  0  0  DC6  4  4  0  0  4.2  10.4  0.8  0  0.4  0  0  0  DC7  0  0  0  0  0  3.2  0  0  0  0  0.2  0  DC8  0.8  0  0  0.8  0  86.4  0  0  0  3.4  0  0  Coloeoptera  CA1  18.4  12.6  7.2  8.8  1  0.8  8.6  6.4  23.6  12.4  5.8  14.4  CA2  0.8  0  0.4  0  0  0  2.2  1.8  2.4  0.2  0.6  1.4  CA3  0  0  0  0  0  0  0  0  0  0  0  0  CA4  0  0  0  0  0  0  0  0  0  0  0  0  CA5  0  0.8  0  0  0  0  0  0  0  0  0  0  CL1  0  0  0  0.2  0  0  0.8  0  0.2  4.4  0  0.4  CL2  3.2  1.6  5.6  16.6  0.8  0.8  8  0.6  3.6  9.8  0.8  15.6  CL3  13.6  0  2.4  4.6  0  0  2.8  0.4  0  3.6  0  6.6  CL4  0  0  0  0  0  0  0  0  0  0  0  0  CL5  0  0  0  0  0  0  0.8  0  0  0  0  0  CL6  0  0  0  0.8  0  0  0  0  0  0  0  0      218     Taxon 13 14  15  16 17 18 19 20 21  22  23 24 Diptera  D1  73  7.4  31.4  12.8  92.4  4.8  25  20  337.4  391  22.2  20.4  D2  2.6  1.6  7.6  1.8  3.2  0  24.4  1  1.6  0.4  0.2  1.4  D3  0  0  0  0  0  0  0  0  0  0  0  2.6  D4  0  0  0  0  0  0.8  0  0.4  0  0  0  0.8  D5  1  0  0  0  0.2  3.6  0.4  0.6  0  0  0  0  D7  0  0  0  0  0  0  0  0  0.8  0  2.6  0  D8  0  0  0  0  1.2  3.2  0  0  2.4  0.4  0  0  D9  0  0  0  0  0  0  0  0  1.6  0  0  0  D10  0  0  0  0.8  0  0  0  1.2  0  0  0  0  D11  0  0  0  0  0  0  0  0  0  0  0  0.4  D12  0  0  0  0  0  0  0  0  0  0  0.2  0.4  DC1  0  4  0  4.6  8.2  1.6  0  0  2.4  7.6  1.2  4  DC2  1  3.2  1.4  1.2  1.6  0  0.8  0.6  0.8  0.6  6  4.4  DC3  1.2  2.2  2  7.6  0  0  1.2  1.8  9.4  2.4  1.4  1.2  DC4  18.4  27.6  13.2  15  14.4  10.4  12.2  15.4  162.4  16.4  12.6  35.6  DC5  1.6  0  0  0  0  0  0  0  0  0.2  0  0  DC6  0  0  1.6  1.8  0  0.8  0  1.2  6.4  1.4  0  0  DC7  0  2.6  0  0  0  0  0  0  0  0.4  0.2  0  DC8  0.8  0  0  0.2  0  0  0  0  0  6.8  0  4.6  Coloeoptera  CA1  17.8  16.4  9.8  9.8  13.6  8.8  1.8  6.6  6.6  4.2  8  3.2  CA2  2  2.6  4.2  1.2  1.6  0  2.6  9  0  4.4  5.4  0.8  CA3  0  0  0.8  1.2  0  0  0  0  0  0  0  0  CA4  0  0  0.8  0.2  0  0  0  0  0  0  0  0  CA5  0  0  0  0  0  0  0  0  0  0  0  0  CL1  4.6  1  0  0  2  0  1  1.8  0  0  0  0  CL2  17.6  0.8  5  2.8  24.8  2.4  0.4  1.6  38.4  26.6  5.6  3.2  CL3  0  0  0.4  3.4  12.8  4.4  0  0.6  0  5.2  2.2  26.6  CL4  0  0  0  0  0  0  0  0  1.8  0  0  0  CL5  0  0  0  0  0  0  0  0  0  1.6  0  0  CL6  0  0  0  0  0  0  0  0  0  0  0  0      219     Taxon 1 2  3  4 5 6 7 8 9  10  11 12 Other  O1  64  6.2  157.6  14.8  4.4  9.6  0  79.4  57.8  91.6  0  14.4  O2  40.8  2.6  0  1.6  0  160.8  2.4  0  1.2  1  0  0.8  O3  2.8  9.6  0  46.4  0  0  46.8  6.6  0  0.8  0.2  9.8  O4  2  0.6  0  1  0.4  12.8  0.2  5.6  0  1.6  0.8  3  O5  0  0.8  0  0.2  0  3.2  0.8  0  0  0  0  0.8  O6  0.8  0  0  0  0  4.4  0  0.2  0.4  0  0  0  O7  1.2  0  0  0  0  0  0  0.8  0  0  0  0  O8  1.2  0.8  1.6  5.6  1.6  0  0  0.8  0.4  0  0  1.4  O9  0  0  0  0.2  0  0  0.4  0  0  0  0  0  O10  0  0  0  0  0  0  0  0  0  0  0  0  O11  0  0  0  0  0  0  0  0  0  0  0  0  O12  0  0  0  0  0  0  0  0  0  0  0  0  O13  0  0  0  0  0  0  0  0  0  0.8  0  0  O14  0  0.2  0  0  0  0  0  0  0  0  0  0  O15  0  0  0  0  0  0  0  0.2  0  0  0  0  O16  0  0  0  0  0  0  0  0  0  0  0  0      220     Taxon 13 14  15  16 17 18 19 20 21  22  23 24 Other  O1  0.8  12  8.4  0.4  59.2  250.4  0.6  44.2  45.8  0  0  10  O2  0  0.8  3.2  2  6.6  3  0  0.2  42  10.4  10.4  22  O3  0  0  22.8  81.8  16.6  3.4  0.6  2.6  0  0  0  0.8  O4  0  0  0  38.8  3.2  20.2  1.8  25.6  0  0  0.2  0.8  O5  0  0  0  0  0  0  0  0  9.6  0  0  0  O6  0  0  0  0  0  0  0  0  7.4  0  0  0.8  O7  0  0  0  0  0.6  0  0  0  0  0  0  0  O8  0.8  0  0  0  9.6  2.6  1.8  2.8  0  0  0  13.4  O9  0  0  0  0  0  0  0  0  0  0  0  0  O10  0  0  0  0.8  0  0  0  0  0  0  0  0  O11  0  0  0  1.8  0  0  0  0  0  0  0  0  O12  0  0  0  0  0  0  0.2  0.8  0  0  0  0  O13  0  0  0  0  0  0  0  0  0  0  0  0  O14  0  0  0  0  0  0  0  0  0  0  0  0  O15  0  0  0  0  0  0  0  0  0  0  0  0  O16  0  0  0  0  0  0  0  0  0  0.2  0  0      221     Appendix  5: Raw  invertebrate data  collected  from  artificial  channels  in  three Hawke’s Bay,  New  Zealand  streams,  December  2009  (Chapter  6).    Values  are  the  pooled  number  of  individuals for the three samples from each channell.    No.  Taxon EX 1. 0. C  EX 1. 0. O   EX 1. 16 .C   EX 1. 16 .O   EX 1. 8. C  EX 1. 8. O   EX 1. 4. C  EX 1. 4. O   Ephemeroptera  1  Deleatidium spp.  15  65  51  70  19  32  9  47  2  Zephlebia dentata  0  0  0  0  0  0  0  0  3  Austroclima sepia  1  0  0  0  0  0  0  1  4  Neozephlebia scita  0  0  0  0  1  0  0  2  5  Coloburiscus humeralis 0  0  0  0  1  0  0  1  Trichoptera  6  Aoteapsyche colonica  42  48  69  48  24  19  18  2  7  Psilochorema A  1  0  0  0  0  0  0  0  8  Hydrobiosis parumbripennis 1  5  0  0  0  1  0  1  9  Hydrobiosis frater  0  1  5  1  0  0  0  0  10  Neurochorema forsteri 0  0  0  4  1  1  0  0  11  Pycnocentria evecta  0  0  0  1  0  0  2  0  12  Pycnocentrodes aeris  7  35  16  41  6  64  19  23  13  Olinga feredayi  0  2  0  0  3  0  0  3  14  Helicopsyche poutini  10  1  41  83  14  12  16  44  15  Confluens hamiltoni  0  0  0  0  0  0  0  0  16  Hudsonema amabile  0  0  0  0  0  0  0  0  Diptera 17  Aphrophila neozelandica 0  0  0  0  0  0  0  0  18  Austrosimulium spp.  0  0  0  0  0  0  1  0  19  Orthocladiinae A  0  2  0  0  0  0  0  0  20  Orthocladiinae B  0  0  0  1  0  2  2  3  21  Orthocladiinae C  0  0  0  0  0  0  0  0  22  Orthocladiinae E  0  0  0  0  0  0  1  1  23  Chironominae A  0  0  0  0  0  0  0  0  24  Chironominae B  0  0  0  0  0  0  0  0  25  Tanypodinae A  0  0  0  0  0  0  0  0  26  Maoridiamesa spp.  0  0  0  0  1  1  0  0  Coleoptera  27  Elmidae Hydora spp.  0  0  0  0  0  0  0  0  28  Hydraenidae  0  0  0  0  0  0  0  0  Mollusca  29  Potamopyrgus antipodarum 15  6  19  33  4  0  4  6  30  Physa spp.  0  0  0  0  0  0  0  0  Other  31  Paracalliope spp.  0  2  1  0  0  0  1  2  32  Platyhelminthes  39  4  58  64  77  16  19  19  33  Oligochaeta  7  4  2  8  0  0  0  2  34  Archichaulioides diversis 0  3  1  0  0  0  0  0  222         No.  EX 2. 0. C  EX 2. 0. O   EX 2. 16 .C   EX 2. 16 .O   EX 2. 8. C  EX 2. 8. O   EX 2. 4. C  EX 2. 4. O   1  1  6  9  8  16  12  1  17  2  2  0  0  0  0  0  0  0  3  0  0  2  0  0  0  0  0  4  0  0  0  0  0  0  0  0  5  0  0  0  0  0  0  0  0  6  10  5  0  0  0  0  0  0  7  6  9  3  1  6  2  8  6  8  4  1  1  0  0  8  0  0  9  0  0  0  0  0  0  0  0  10  0  0  0  0  0  0  0  0  11  8  2  3  1  0  0  0  0  12  26  44  4  17  14  23  18  12  13  0  0  0  0  0  0  4  0  14  0  0  0  0  0  0  0  0  15  0  1  0  0  0  0  0  0  16  0  0  0  0  0  0  0  0  17  0  0  0  0  0  0  0  0  18  2  0  3  0  0  0  0  0  19  0  0  0  0  0  4  0  0  20  12  7  1  3  3  8  21  12  21  2  0  0  0  0  0  4  8  22  0  0  0  0  0  0  0  0  23  1  0  0  0  1  0  0  4  24  0  0  0  2  0  0  0  0  25  0  0  0  0  0  0  8  4  26  0  0  0  0  0  0  0  0  27  3  5  1  1  6  1  0  0  28  0  0  0  0  0  0  0  0  29  78  82  247  182  249  240  306  588  30  0  0  0  0  1  0  1  0  31  0  0  11  0  3  3  51  36  32  2  0  0  1  4  6  5  8  33  5  6  27  13  419  172  55  45  34  0  1  0  0  0  0  0  0  223         No.  EX 3. 0. C  EX 3. 0. O   EX 3. 16 .C   EX 3. 16 .O   EX 3. 8. C  EX 3. 8. O   EX 3. 4. C  EX 3. 4. O   1  1  8  3  12  6  15  1  5  2  0  0  0  0  0  0  0  0  3  0  0  0  0  0  0  0  0  4  0  0  0  0  0  0  0  0  5  0  0  0  0  0  0  0  0  6  16  0  2  0  0  1  0  0  7  11  6  3  2  4  3  2  6  8  0  0  0  0  0  0  0  0  9  0  0  1  0  0  0  0  0  10  0  0  0  0  0  0  0  0  11  0  0  0  0  0  0  0  1  12  13  47  2  47  16  30  16  42  13  0  0  0  0  0  0  1  0  14  0  0  0  0  0  0  0  0  15  0  0  0  0  0  0  0  0  16  0  1  0  4  0  1  0  8  17  0  1  0  1  1  1  0  0  18  5  1  4  1  1  0  1  0  19  0  0  0  0  0  0  0  0  20  1  0  0  1  0  2  1  3  21  0  0  0  0  0  1  1  0  22  0  0  0  0  0  0  0  0  23  0  0  0  0  0  0  0  0  24  0  0  0  0  0  0  0  0  25  0  0  0  0  0  0  0  0  26  0  0  0  0  0  0  0  0  27  0  0  0  0  0  0  0  0  28  0  0  0  1  0  0  0  0  29  25  57  89  122  49  94  15  158  30  0  0  0  0  0  0  0  0  31  0  0  0  4  2  2  1  4  32  4  0  0  1  2  0  0  0  33  16  7  11  20  38  24  17  28  34  0  0  0  0  0  0  0  0    224     Appendix 6: Reprint of co‐authored paper in New Zealand Journal of Ecology.   Collier, K.  J., B. Aldridge, B.  J. Hicks,  J. Kelly, A. Macdonald, B.  J. Smith, and  J. Tonkin. 2009.  Ecological values of Hamilton urban streams (North  Island, New Zealand): constraints  and opportunities for restoration. New Zealand Journal of Ecology 33:177‐189.    225       226       227       228       229       230       231       232       233       234           235       236           237     Appendix  7:  ‘Early  view’  reprint  of  co‐authored  paper  published  in  River  Research  and  Applications.    Schwendel, A. C., R. G. Death, I. C. Fuller, and J. D. Tonkin. In press. A new approach to assess  bed stability relevant for invertebrate communities in upland streams. River Research  and Applications.        238         239         240         241         242         243         244         245         246         247         248         249         250     251