UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Seasonal food limitation of detritivorous insects in a montane stream Richardson, John Stuart 1989

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

Item Metadata

Download

Media
831-UBC_1989 A1 R52.pdf [ 7.1MB ]
Metadata
JSON: 831-1.0106745.json
JSON-LD: 831-1.0106745-ld.json
RDF/XML (Pretty): 831-1.0106745-rdf.xml
RDF/JSON: 831-1.0106745-rdf.json
Turtle: 831-1.0106745-turtle.txt
N-Triples: 831-1.0106745-rdf-ntriples.txt
Original Record: 831-1.0106745-source.json
Full Text
831-1.0106745-fulltext.txt
Citation
831-1.0106745.ris

Full Text

S E A S O N A L F O O D L I M I T A T I O N O F D E T R I T I V O R O U S I N S E C T S I N A M O N T A N E S T R E A M B y J O H N S T U A R T R I C H A R D S O N B.Sc.(Honours), The University of Toronto, 1979 M . S c , The University of Alberta , 1983 A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F D O C T O R O F P H I L O S O P H Y i n T H E F A C U L T Y O F G R A D U A T E S T U D I E S (Department of Zoology) We accept this thesis as conforming fT/dHhe, required standard T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A July 1989 © John Stuart Richardson, 1989 National Library Bibliotheque Rationale of Canada du Canada Canadian Theses Service Service des theses-canadiennes Ottawa, Canada K1 A 0N4 The author has granted an irrevocable non-exclusive licence allowing the National Library of Canada to reproduce, loan, distribute or sell c o p i e s of his/her thesis by any means and in any form or format, making this thesis available to interested persons. The author retains ownership of the copyright in his/her thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without his/her per-mission. L'auteur a acc o r d e une li c e n c e irrevocable i non exclusive permettant a la Bibliotheqi nationale du Canada de reproduire, prete distribuer ou vendre d es c o p i e s de s a thej de quelque maniere et sous quelque forrr que c e soit pour mettre des exemplaires c cette these a la disposition d e s person™ inte r e s s e e s . L'auteur conserve la propriete du droit d'aute qui protege s a these. Ni la these ni des extrai substantiels de celle-ci ne doivent et imprimes ou autrement reproduits sans S( autorisation. ISBN 0-315-55188-7 Canada In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. (Signature) Department of Zoology The University of British Columbia Vancouver, Canada Date 26 July 1989 DE-6 (2/88) Abstract M a n y species of stream invertebrates gain most of their energy by the consump-tion of coarse detrital materials. While most of these organisms are univoltine or semivoltine, the biomass of assimilable detritus varies seasonally as a result of sev-eral processes. The period of detritus input is highly seasonal, decomposition rates are positively temperature dependent, and winter spates result i n fragmentation and flushing of detrital materials. From two years of measuring detritus inputs and standing crops, I showed that the abundance of this resource varied by almost two orders of magnitude seasonally. Since many consumers which rely on this resource have generation times equal to that of the period of resource variation, individuals and perhaps populations may be faced with periods of low food abundance. This work addressed the consequences of seasonal food l imitation of stream insects. To test this food Limitation hypothesis, I experimentally manipulated detritus input rates to otherwise natural communities of stream benthos using a replicated, 3 - treatment design. These experiments were conducted i n experimental streams i n the University of Br i t i sh Columbia research forest over the course of one year. Increasing input rates of detritus resulted i n large increases i n size at maturity and growth rates for 7 of 9 common species. This was true for both summer and winter emerging species. Increased supply of detritus also resulted i n increased densities and higher rates of colonization for some species. There was no evidence of change i n phenology for any species. The densities of the chironomid Brillia retifinis (the only species studied that had a short generation time) underwent exponential growth during the first 3 months of the experiment, reaching densities 10x those of the control and natural streams. This species apparently fills the role of a "fugitive" i n this system. One alternative hypothesis for increased densities following addition of whole leaf detritus was a significantly altered microhabitat. To test this possibility I i i compared the use of real and artificial (polyester) leaf packs by stream invertebrates. Those species which typically consume coarse detritus were almost never found on the artificial leaf packs, while they attained high densities on the real leaves. In contrast, fine-particle, and algae consuming species were found i n similar densities on artificial and real leaf packs, although there was a time lag i n colonization of the polyester leaves. These results suggest that microhabitat alone cannot lead to increased densities of detritivores. The densities of species which do not consume large particles of detritus also were affected by whole-leaf additions. Density of total consumers of fine parti-cles of detritus increased when coarse detritus was supplemented and most taxa showed this response. This result was apparently an indirect effect of diminution of detrital particle size by larger detritivores. Predaceous species also increased i n density under detritus supplementation. Increased densities of taxa other than large-particle-detritus feeders indicates that effects at one trophic level can affect other trophic levels. i i i Table of Contents Page Abstract 1 1 List of Tables vii List of Figures viii Acknowledgements xi Chapter One Introduction 1 The stream system 3 Peculiarities of detritus as a resource 5 The problem - thesis outline 6 Chapter Two Detritus dynamics in small montane streams 8 Study site 9 Methods and Materials 11 Direct inputs 11 Benthic detritus standing crops 12 Processing rates 12 Results 13 Direct litterfall 13 Benthic detritus 14 Decomposition rates 20 Discussion 22 Chapter Three Natural history of stream detritivores 29 Materials and Methods 29 Results 30 Discussion 37 Chapter Four Seasonal food limitation of detritivores in a montane stream: an experimental test 42 Materials and Methods 44 i Statistics 48 iv Results 49 Life cycles 49 Benthos 49 Comparisons with natural stream 49 Init ial conditions 50 Responses to food manipulations 50 Immigration and emigration 55 Adul t mass at emergence 60 Summer-emerging species 64 Winter-emerging species 70 Biomass and adult numbers 72 Other species 75 G u i l d structure 77 Discussion 77 Food quantity and quality 85 Abiot ic conditions 87 Competit ion, coexistence and community structure 87 C h a p t e r F i v e Indirect effects of detritus manipulations 92 Materials and methods 93 Statistics 95 Results 95 Benthic densities 95 Emigration and immigration 100 Adults 105 Fine particulate organic matter 105 Discussion 107 C h a p t e r Six Food, habitat, or both? : macroinvertebrate use of leaf accumulations in streams I l l Methods and materials 112 Study site 112 Leaf pack manipulations 113 Results 115 v Leafpack decomposition rates 115 Macroinvertebrates 116 Colonization 123 Discussion 127 Chapter Seven General discussion 131 Implications for community structure 132 Evolution of life cycle phenology 133 Future research directions 134 Literature Cited 137 Appendix 1 152 vi List of Tables Table I — Summary of stream characteristics at the three study sites 10 Table II — Ash-free dry mass of components of direct litterfall to small coastal streams 16 Table III — Summary of the streams and dates of the alder leaf pack decomposition trials 21 Table IV — Summary ANCOVA of decomposition rates of red alder leafpacks .22 Table V — Summary of life cycles of common detritivores in Mayfly Creek, BC. 31 Table VI — List of size and mass equations for each taxon for use in calculating biomass from head width measurements 35 Table VII — Summary nested ANCOVA tables for mass at emergence of adults from experimental streams 67, 68 Table VIII — Percentage change in adult dry mass relative to adults from the "natural" treatment 69 Table IX — Results of tests of treatment effects on benthic densities of invertebrate taxa for two different periods 97 Table X — Results of tests of treatment effects on emigration or immigration rates of invertebrate taxa 104 Table XI — Summary of ANCOVA tests of treatment effects on adult insects cap-tured in emergence traps, controlling for sex and date of emergence 106 Table XII — Summary ANCOVA table for decomposition rates of alder leaf packs, comparing the rates in riffle and pool areas 116 Table XIII — Summaries of ANOVA's for numbers per leaf pack 118 vn L i s t o f F i g u r e s Figure 1 — Dai ly litterfall rates for all detritus collected i n litter traps along the margins of U B C Research Forest streams 15 Figure 2 — Dai ly litterfall rates for deciduous leaves collected i n litter traps along the margins of U B C Research Forest streams 17 Figure 3 — Dai ly litterfall rates for coniferous leaves collected i n litter traps along the margins of U B C Research Forest streams 18 Figure 4 — Estimates from Mayfly Creek of benthic standing crops from August 1985 to July 1986 19 Figure 5 — Relationship between rate of loss per day for leaf packs of red alder versus the average temperature during the incubation period 23 Figure 6 — Seasonal development patterns of Zapada cinctipes, Zapada haysi and Capnia spp. i n Mayfly Creek, Br i t ish Columbia 32 Figure 7 — Seasonal development patterns of Malenka species and Despaxia augusta i n Mayf ly Creek, Br i t ish Columbia 34 Figure 8 — Seasonal development patterns of Lepidostoma roafi and Brillia retifinis i n Mayfly Creek, Bri t ish Columbia 36 Figure 9 — Patterns of geographic variation i n the t iming of the first date of emer-gence of Zapada cinctipes summarized from literature reports 38 Figure 10 — Rates of leaf litter input to stream channels for the three manipulative treatments 46 Figure 11 — Benthic densities of detritivores i n Mayfly Creek, the instream chan-nels, and the streamside channels 51 Figure 12 — Benthic densities of common detritivores from streamside channels following treatment by additions of whole leaf detritus 53, 54 Figure 13 — Numbers of emigrant larvae of Malenka spp. and Lepidostoma roafi from the streamside experimental channels 57 Figure 14 — Numbers of emigrant larvae of Zapada spp. and Brillia retifinis .. 58 Figure 15 — Daily per capita emigration rates of Brillia retifinis 59 vi i i Figure 16 — Dai ly colonization rates of detritivores i n the streamside experimental channels (immigration — gross emigration) 61, 62 Figure 17 — Head widths (geometric means) of Zapada cinctipes nymphs for emi-grants (filled symbols) and benthic individuals (open symbols) for each treat-ment 63 Figure 18 — Mass at emergence (mg dry mass) for common summer emerging detritivores from the experimental streams caught i n emergence traps . 65, 66 Figure 19 — Mass at emergence (mg dry mass) for three winter stoneflies, showing least square means for treatment 71 Figure 20 — Seasonal changes i n benthic standing crop of all nine detritivore taxa considered 73 Figure 21 — Numbers of adults collected from emergence traps for each stream plotted against mean food input rate for A . summer species and B . winter species 74 Figure 22 — Combined biomass of adults collected for emergence traps for all channels, wi th summer and winter species plotted separately, versus leaf input rate 76 Figure 23 — Relative detritivore guild composition by percentage biomass 78 Figure 24 — Densities of four taxa of collector which were significantly different among treatments 98, 99 Figure 25 — Densities of total collectors, predators and grazers i n experimental stream channels 101 Figure 26 — Species richness measured as taxa per sample 102 Figure 27 — Regressions of alder leaf pack mass remaining over time i n Mayfly Creek 117 Figure 28 — Malenka nymphs per leaf pack, alder and artificial, over period of residence in situ 120 Figure 29 — Lepidostoma roafi nymphs per leaf »pack, alder and artificial, over period of residence in situ 121 Figure 30 — Numbers of large (> 1.0 mm) detritivores per leaf pack, over period of residence in situ 122 ix Figure 31 — Numbers of Paraleptophlebia debilis and P. temporalis per leaf pack 124 Figure 32 — Numbers of Baetis spp. per leaf pack 125 Figure 33 — Numbers colonizing leaf packs within two days of placement into stream 126 x A c k n o w l e d g e m e n t s M a n y people helped out at various stages i n the completion of this study and to whom I am grateful; unfortunately inclusion of al l their names would be i m -possible. To single out several individuals I wish to thank Brad Anhol t , David Bernard and B i l l Nei l l , all of whom gave generously of their t ime, energy, insight and friendship. B i l l Nei l l has been a tremendous supervisor and I appreciate his faith i n me and all his efforts on my behalf. M y thanks go to the rest of my super-visory committee: T . H . Carefoot, J . H . Myers, T . G . Northcote and A . R . E . Sinclair. I've enjoyed my association with the Ecology Group (formerly the Institute of A n -imal Resource Ecology), all the discussions and seminars, and especially their help and friendship. The staff of the Malcolm Knapp Research Forest were helpful i n many ways during the course of field work. Taxonomic identifications or verifica-tions were provided by Richard Baumann, Syd Cannings, Andrew Nimmo, Donald Oliver, John Weaver III, Er ic W h i t i n g and Glenn Wiggins. Thanks also to my ex-ternal examiners, Drs. D . D . Hart and J .R . Webster for their comments. This study was funded through an N S E R C Canada operating grant to D r . W . E . Nei l l . Per-sonal support was provided through scholarships from N S E R C and The University of Br i t ish Columbia. M y bride Daphne deserves special recognition for having stood by me and encouraged me through a decade as a student; she's been instrumental i n any success I may have achieved. x i Chapter One INTRODUCTION Resource l imitation is defined as a l imit to population growth rate imposed by any resource i n short supply (Sinclair in press) or potentially i n short supply if there are contests for resources. The term l imitation is used here i n contrast to regulation which requires some form of density dependence that provides rapid demographic feedback, i n turn damping population fluctuations. L imita t ion therefore makes no assumptions about density dependence but does not imply that a l imit ing resource cannot also be regulating. For example, food supply can be l imi t ing , but also regulating because the feedback between food production and consumption is often rapid (Neil l 1975, White 1978, Emlen 1984). In some systems the lag time i n feedback can be destabilizing (May 1981). Resource hmitation is a necessary condition for competition (Ti lman 1982, Schoener 1983) and is a central assumption of many models of territoriality and evolution by natural selection. Given the importance of the assumption of l imit ing resources to these tenets it is surprising for how few systems the role of resource hmitation is understood. There are several kinds of data that are often inter-preted as evidence for food hmitation: (1) positive correlations among densities or productivity of consumers and their food resources, (2) negative correlations be-tween density and average individual biomass, and (3) energetic supply-demand budgets approaching a value of unity (Neill 1975, Eadie and Keast 1982, Rough-garden 1986). B y far the strongest test of a food Hmitation hypothesis involves manipulative experiments, with food supply or consumer densities as the treatment variables (Roughgarden 1986). Several of the issues i n ecology implicate the role of food i n l imit ing population density. One of these debates concerns the relative importance of resource l imita-tion versus disturbance as influences on population size. The latter of these two 1 perspect ives suggests tha t p o p u l a t i o n densit ies are r e d u c e d b y d i s t u r b a n c e to levels at w h i c h resources r a r e l y become l i m i t i n g ( A n d r e w a r t h a a n d B i r c h 1954). A second issue concerns the roles of c o m p e t i t i o n versus p r e d a t i o n i n l i m i t i n g p o p u l a t i o n size (Schoener 1982, 1983, C o n n e l l 1983). If predators l i m i t p o p u l a t i o n d e n s i t y t h e n re-sources cannot be l i m i t i n g a n d vice-versa. T h i r d , one m o d e l of c o m m u n i t y f u n c t i o n ( H a i r s t o n et al. 1960), a n d la ter e laborat ions ( O k s a n e n et al. 1981, F r e t w e l l 1987), suggests tha t ent ire t r o p h i c levels s h o u l d be l i m i t e d ei ther b y f o o d or predators . W h i c h fac tor l i m i t s a t r o p h i c leve l depends o n the n u m b e r of t r o p h i c l i n k s a n d habi ta t p r o d u c t i v i t y , b u t the m o d e l predic ts tha t adjacent t r o p h i c levels have alter-nate l i m i t i n g factors . It also predic ts tha t detr i t ivores s h o u l d be f o o d l i m i t e d . A l l of these issues i n d i c a t e that f o o d m a y be a n i m p o r t a n t l i m i t i n g fac tor for n a t u r a l p o p u l a t i o n s . T h e resource s u p p l y of m a n y organisms varies seasonally (def ined as a n y p e r i -o d i c p h e n o m e n a w i t h a one year p e r i o d ) a n d , i f resources are i n short s u p p l y , the i n t e n s i t y of c o m p e t i t i o n a n d d e m o g r a p h i c rates s u c h as s u r v i v o r s h i p m a y also v a r y seasonally. S h o r t - l i v e d organisms ( < 1 year) w i l l exh ib i t large seasonal changes i n p o p u l a t i o n size ( M c C a u l e y a n d M u r d o c h 1987, W a l t e r s et al. 1987). F o r organisms w i t h l i fespans of a year or longer these changes i n resource s u p p l y take place w i t h i n a generat ion a n d i t is less obvious h o w changes i n resource s u p p l y affect s u c h p o p -ula t ions (Fre twel l 1972). Responses to seasonal changes i n resources c o u l d i n c l u d e d o r m a n c y , m i g r a t i o n , changes i n d e a t h , b i r t h or g r o w t h rates , or a n y c o m b i n a t i o n . Seasonal i ty is a c o m m o n feature of most env i ronments ( F r e t w e l l 1972, C o l w e l l 1974, W o l d a 1988). A n n u a l p e r i o d i c i t y has o b v i o u s l y h a d a n i m p o r t a n t role i n or-ganic e v o l u t i o n . S u c h c o m m o n l i fe h i s t o r y features as m i g r a t i o n , h i b e r n a t i o n , t i m i n g of r e p r o d u c t i o n a n d diapause , i l lus t ra te some of t h e r e s p o n s e s to seasonali ty. Sea-sona l i ty can occur i n m a n y forms , such as t empera ture , p h o t o p e r i o d , p r e c i p i t a t i o n pat terns , a n d m a y be m e d i a t e d t h r o u g h resources w h i c h r e s p o n d to e n v i r o n m e n t a l 2 patterns or by interactions with other species. Seasonal changes i n the environment are also predictable i n general way. The problem of incorporating seasonality into models of population dynamics has hindered progress towards determining the consequences of periodic environ-ments for populations. Most models which have attempted to address the prob-lem have been l imited to age-structured or continuously reproducing populations. Models for age-structured populations have considered the year as divided into two discrete periods, one favourable, the other less so (Fretwell 1972, Hastings 1984). For populations with continuous reproduction the problem is more straight forward; seasonality is usually portrayed as a function of resources, often following a sine wave (Nisbet and Gurney 1976, Boyce and Daly 1980, Cushing 1986). Thus far, models for within-generation changes of univoltine populations with non-overlapping co-horts are undeveloped, although Brown (1989) produced a model for coexistence of two species on a single, seasonal resource based on a trade-off of foraging efficiency and dormancy at different resource densities. The stream system In the majority of small streams (orders one to three) allochthonous inputs of organic carbon, i n the form of terrestrial vegetation, constitute the main source of biologically available food energy (Nelson and Scott 1962, Minshal l 1967, Hynes 1970, Fisher and Likens 1973). Because small streams are covered by a canopy of surrounding terrestrial vegetation, in-stream photosynthesis is Hixdted by shading, and fallen terrestrial plant matter is the major organic carbon source (Hynes 1975). Whi le there are many exceptions to this general pattern, most systems show trends that are consistent with it (Minshall et al. 1983, Naiman et al 1987). Community consequences of this generality are a large proportion of detritivorous individuals or species and a low ratio of primary production to respiration (Vannote et al. 1980, Minshal l et al. 1983). 3 There are many species of lotic invertebrates which are exclusively detritivorous (or almost so) during some stage of their life cycles (Cummins 1973). For several reasons it has been useful to distinguish detritivores on the basis of size of detrital particles consumed. Among these reasons are differences i n the way food is handled, e.g., species feeding on small particles (< 1.0 mm) generally collect many particles simultaneously using fan or brush-like appendages, or silk or mucous nets to trap food; these species have been referred to as collectors (Cummins 1973). Species consuming large particles of detritus (> 1.0 mm) generally handle one particle at a time, have chewing mouthparts and are referred to as shredders (Cummins 1973, Merritt and Cummins 1984). The microbiota also vary with particles of different sizes, w i t h large particles primarily consumed by fungi and small particles by bacteria (Cummins and K l u g 1979, Cummins et al. 1989). There are several sources of seasonal variation i n the standing crop of coarse particulate organic matter i n streams ( C P O M - detritus particles > 1.0 m m diame-ter). Input of C P O M occurs primarily during the autumn at the time of abscission of leaves from riparian vegetation. Rates of decomposition are positively correlated wi th temperature ( M c A r t h u r et al. 1988), so that faster decay coincides wi th lower detrital input rates and generally higher metabolic requirements of invertebrate detritivores. As a result of these processes, C P O M abundance can vary up to 2 orders of magnitude seasonally (Minshall 1967, Short and W a r d 1981, Barlocher 1983, Webster et al. 1983, Petersen et al. 1989). The C P O M remaining i n the stream after the autumnal peak of input becomes increasingly composed of refrac-tory, highly lignified compounds, and its soluble protein may decline by 50% by spring (Barlocher 1983). There are many other aspects to seasonality i n streams and these may interact i n complex ways. For example, seasonal cycles of temperature and light affect many processes i n streams, such as algal productivity, microbial growth and respiration, 4 animal growth, and timing of development. In coastal regions of western North America, precipitation patterns are distinctly seasonal wi th wet autumns and win-ters and relatively drier summers (Feller and Kimmins 1979). In terms of stream discharge, a seasonal pattern is obvious with frequent freshets during the cooler part of the year, and relatively stable and low base flow during summer (Shortreed and Stockner 1983, Feller and Kimmins 1979). Peculiarities of detritus as a resource One important aspect of this detritus - detritivore system is that the dynamics are donor-controlled. The essential difference from normal food chains is that a feedback loop from consumers to producers does not exist since detritivores cannot affect the rate at which the donors deliver resources. Lotic detritivores cannot feed directly on the riparian vegetation and thus cannot affect the resource renewal rate. The dynamics of detritus-detritivore systems i n isolation from the rest of the community are considered to have very high stability (DeAngelis 1975, P i m m 1982). Most animals are incapable of util izing the majority of the carbon i n plant detritus, much of which is cellulose or bound by tannins (Mar t in et al. 1980, 1981). Some animals, such as termites, get around this by carrying mutualist microflora i n their guts. Most animals without such symbionts have a gut p H too low to break the strong binding of tannins, and instead rely on microbes associated with detritus to begin the conversion of detrital carbon to more easily assimilable products (Mart in et al. 1980). These products are the microbes themselves and/or substrates partially "digested" by the activities of microbes (Arsuffi and Suberkropp 1988). Throughout this thesis I w i l l use the term detritivore to refer only to those lotic species which feed on detrital particles > 1.0 m m during part of their life cycle. These taxa are frequently referred to as "shredders" i n the literature, following Cummins (1973). However, while Cummins' term is descriptive of the feeding style of many coarse-particle detritivores, none of the C P O M consumers I have explicitly 5 considered i n this thesis shred whole particles, but instead feed by scraping the leaf matrix from among the vascular bundles (described by Wallace et al. 1970). The problem and thesis outline During the past 30 years, research on detritus-detritivore systems has concen-trated on mineralization and energy flow through this pathway. This has been the primary focus i n streams (Cummins et al. 1983, B i r d and Kaushik 1981, Web-ster and Benfield 1986), lakes, salt marshes, and terrestrial systems (Dickinson and Pugh 1974, Swift et al. 1979). In addition, there has been much work examin-ing consumer choice, growth rates and conversion efficiencies on different types of detritus (Barlocher and Kendrick 1973, 1975, Rietsma et al. 1988). In contrast, the way that quantity and quality of detritus affect populations of detritivores has received very litt le consideration. In "cafeteria"-style preference experiments most aquatic invertebrates exhibit distinct choices of detritus based on type or degree of fungal or microbial conditioning. Most taxa have a similar rank order to their preference of particular detrital types, e.g., deciduous leaves > coniferous leaves > wood. Whi le invertebrate preference also extends to different types of deciduous leaves, most consumers w i l l eat what is off erred i n the absence of choice. Thus, it seems reasonable to consider deciduous leaf detritus as a single resource of variable quality for detritivores. The assumption that leaf detritus is a singular, poorly dif-ferentiated resource simplifies consideration of seasonal variability i n quantity and quality. Given the sources of seasonal variation i n abundance and quality of C P O M outlined above, food supply may limit population growth of lotic detritivores. The hypothesis that stream detritivores may be food l imited has been proposed previ-ously (Anderson and Sedell 1979:361, Shiozawa 1983), but has never been tested experimentally. M a n y other observations also suggest that populations of stream detritivores are food l imited, and these are reviewed i n Chapter four. 6 This thesis addresses the problem of seasonal variation i n resource abundance and the consequences i n terms of food l imitation for detritivorous insects i n a small, coastal, montane stream i n Bri t ish Columbia. In the second chapter I detail a num-ber of components of the detrital resource base i n this small stream as a way of defining the availability of food to these detritivorous insects. In chapter three I outline features of the life cycles of the common species dealt wi th i n this system. The fourth chapter describes a food supplementation experiment designed to test the hypothesis that detritivores are seasonally food l imited i n this system. Chap-ter five examines the responses of non-detritivore species to the manipulation of detrital input rates during the food supplementation experiment. The sixth chap-ter describes an experiment to examine use of coarse detritus as a component of habitat by stream insects, i n order to rule out microhabitat alteration as a cause of increased detritivore densities i n streams treated with added food. The final chapter provides an overview, some speculation on the evolutionary importance of seasonal food l imitation, and prospects for future research on the topic. 7 C h a p t e r T w o DETRITUS DYNAMICS IN SMALL MONTANE STREAMS Small streams receive most of their energy inputs from allochthonous materials (i.e., from outside the system), mostly i n the form of detritus from riparian vegeta-tion (reviewed by Anderson and Sedell 1979, Cummins et al. 1983). Some estimates indicate that greater than 60% of the energy of stream community metabolism is contributed by allochthonous detritus (Nelson and Scott 1962, Fisher and Likens 1973, Triska et al. 1982, Connors and Naiman 1984). The importance of this mate-rial to biological processes i n streams makes it vital to understand the pathways and rates of both input and decomposition of detritus. The seasonal t iming of inputs and outputs of detrital materials are thought to be especially important to the life cycle phenology of species which consume detritus, particularly those species which are obligate detritivores (Petersen and Cummins 1974). Different types of detritus are distinguished by consumer species and should be considered separately for this and the following reasons. First , the rates of decomposition of leaf detritus from different deciduous plants alone vary at least 16-fold depending upon source (Webster and Benfield 1986). Conifer needles and woody material decay at rates much slower than deciduous leaf detritus (Sedell et al. 1975, Anderson et al. 1978, Melil lo et al. 1983) and apparently play a l imited role i n the nutrit ion of stream detritivores (Anderson et al. 1978, Webster and Benfield 1986). Second, the nutritional value of detritus varies as a function of type (Barlocher and Kendrick 1973, Sweeney and Vannote 1986). T h i r d , the t iming and rate of detrital input also differs between sources. Final ly, the size and structure of detritus can affect other stream processes, e.g., large wood debris contributes to detrital storage and substrate stabilization (Triska and Cromack 1980). 8 In this chapter I describe aspects of the dynamics of coarse particulate organic matter ( C P O M , detritus > 1.0mm i n diameter) i n three small , coastal montane streams of southwestern Bri t ish Columbia. There are few estimates of detritus abundance i n small coniferous forest streams (Neaves 1978, Naiman and Sedell 1979, Minshal l et al. 1983, Triska et al. 1984), and most estimates make no distinction among types of detritus, e.g., wood versus non-wood. Work done on continental streams have shown large differences i n the seasonal abundance of C P O M , of up to two orders of magnitude (Minshall 1967, Barlocher 1983, Webster et al. 1983). In addition different detrital types vary i n the timing and magnitude of their seasonal abundance i n these streams. I have focussed on C P O M for two reasons. First , while C P O M dynamics are complex (Cummins et al. 1983), they are much clearer than F P O M (fine particulate organic matter) dynamics. Second and most importantly, coarse detritus constitutes the food supply for a number of species of aquatic invertebrates. A s mentioned i n Chapter 1, many detritivores feed on coarse detritus exclusively (or almost so) for much of their larval life. S T U D Y S I T E Three second order streams i n the Malcolm K n a p p Research Forest of the University of Bri t ish Columbia ( U B C - R F ) were studied from 1984 to 1986. The U B C - R F is located i n the Coast Range mountains approximately 60 k m east of Vancouver (122°34' W x 49° 16' N ) . The watersheds lie i n the Coastal West-ern Hemlock forest, and the conifers Tsuga heterophylla (western hemlock), Thuja plicata (western red cedar) and Pseudotsuga menziesii (Douglas fir) are the dom-inant forest species. The major riparian species are Alnus rubra (red alder), Acer circinatum (vine maple) and Rubus spectabilis (salmonberry), based on litter-fall estimates. A l l three streams drain over thin soils of glacial origin, overlying a predominantly acid igneous bedrock (Feller and Kimmins 1979). 9 These streams were chosen as representative of small streams i n the area and also for ease of access. Some of the essential features of the three streams are outlined i n Table I. Blaney Creek is the only stream of the three with a lake upstream of the sampling reach; Plac id Lake is about 2 k m upstream. Mayfly Creek drains the west side of M t . Blanchard and has a steeper slope than the other two streams. Spring and Blaney Creeks drain more gently sloping parts of the forest. T a b l e I — Summary of stream characteristics at the three study sites. Slope (m • m _ 1 ) is shown for a 1 k m length around the study site as approximate slope. Stream Drainage Approximate Slope i n Forest Alt i tude Area km2 Slope Sample Reach Age (m a.s.l.) Blaney 2.9 0.054 0.025 30 380 Mayfly 3.6 0.080 0.038 60 315 Spring 3.8 0.037 0.024 250+ 135 The substrate i n all three streams was similar, made up predominantly of cobble, gravel, and sand. Boulders were common i n sections of Mayfly Creek but were infrequent i n the study section of Blaney Creek and were lacking i n the area of study i n Spring Creek. The forest i n each drainage basin has had a diverse history of logging and fire. The forest age given i n Table I is that of the riparian vegetation surrounding the sampling reach. Mayfly Creek drainage has second growth forest of western hemlock and cedar. The watershed was logged early this century, and consequently it is relatively even aged. Portions of the watershed of Blaney Creek have been logged at different times since 1958, but most of the forest close to the stream was logged and replanted i n 1958. The riparian zone around Spring Creek is old-growth forest, but the majority of the watershed has been logged over the past 110 years. 10 Some of the watershed was logged for experiments on stream hydrology and forest nutrient budgets upstream of the study reach (Feller 1977, Feller and Kimmins 1979). Details of water chemistry and underlying geology for similar streams i n the U B C Research Forest have been published previously (Feller 1977, Feller and K i m -mins 1979). The concentrations of most major ions vary with season and discharge. Average annual p H is 6.8, with electrical conductivity of 20.0/xmho/cm at 25 C . The average annual concentration of PO4 was 10 fj,g • L - 1 , nitrate 230 fig • L - 1 , and C a 1.5 m g - L - 1 (Feller 1977, Feller and Kimmins 1979). Mayf ly and Spring Creeks are generally shaded by dense riparian vegetation so that stream temperatures vary at most 3 C daily. The extremes of temperatures recorded from 1983-1988 were 17.5 C to below 0 C for Mayfly Creek, 17.5 C to 1 C i n Spring Creek, and 26 C to 0 C i n Blaney Creek. Discharge varies considerably seasonally and is characterized by high and variable winter flows and lower, more stable discharge i n summer and early autumn (Feller and Kimmins 1979). M E T H O D S A N D M A T E R I A L S Direct Inputs Litterfal l data were collected using basket traps of aluminum screening (mesh size 1.4 mm) with walls of about 4 cm high and a bottom area of 975 c m 2 . Ten of these traps were placed at approximately 10 m intervals along the sides of each of the three streams. The traps were placed on boulders or logs within the bankfull perimeter of the stream to sample litter that would fall directly into the streams or be entrained by wind movement or by the flow as discharge increased. Traps were occasionally disturbed or washed away, and these were replaced on subsequent visits. Litter traps were emptied at irregular intervals, with samples taken more frequently as rates of deposition increased. Litterfall samples were collected from 31 July 1984 to 3 August 1986. 11 In the laboratory samples were separated into the following groups: i) red alder (Alnus rubra), i i ) vine maple (Acer circinatum), i i i ) Rubus species (R. spectabilis and R. ursinus), i v ) willow (Salix spp.), v ) other deciduous leaves, which i n -cluded sitka alder (Alnus sinuata), dogwood (Cornus nuttallii), and huckleberry (Vaccinium spp.), v i ) cedar leaves (Thuja plicata), v i i ) hemlock (Pseudotsuga het-erophylla) and douglas fir needles (Pseudotsuga menziesii), v i i i ) wood, and ix ) miscellaneous plant material such as flower parts, bud scales, lichens, and moss. Each component was dried for at least 24 hr at 60 C , weighed, reduced to ash i n a muffle furnace at 500 C , then reweighed to determine ash-free dry mass ( A F D M ) . Benthic Detritus Standing Crops Samples of benthic detritus were collected monthly using a cylindrical sampling device of sheet metal wi th a net on the downstream side with mesh size 63/xm. After the macroinvertebrates were removed, the remaining material was sorted by sieve, and materials larger than 5 m m diameter were sorted into deciduous leaves, cedar, hemlock or douglas fir needles, wood, and a category for miscellaneous detritus. Detritus 1 to 5 m m diameter was treated as one category. A l l detritus from benthic samples was expressed as ash-free dry mass. Processing Rates Decay rates of vegetable detritus were estimated using leaf packs of red alder. The leaves for this study were collected shortly after abscission, air dried, heat-sealed i n plastic bags, and stored in a deep-freeze (-20 C) unt i l needed. The leaves were weighed into groups of 5 g (±0 .025 g), fastened together using pieces of monofila-ment nylon and then loosely tied to concrete bricks as described by Petersen and Cummins (1974). Fifteen to 25 leaf packs were placed i n either Mayfly Creek or Spring Creek on different dates and randomly chosen subsets were retrieved on 5 different dates from each series. The first subset was collected two days after placement to serve as an init ial weight, controlling for losses due to handling and 12 leaching, the latter of which results i n rapid loss of mass (Webster and Benfield 1986). The only exception to this pattern was for the 1987 series where leaf packs were placed i n the stream on five different occasions and retrieved simultaneously 2 days after the last set was placed (see Chapter 6). In the laboratory, leaves were rinsed twice to remove macroinvertebrates. A l l leaf fragments > 1 m m were added to the sample which was dried, weighed, ashed, and reweighed as described above. These data produced one decay rate curve for each series of leaf packs. Temperatures were measured with maximum-minimum thermometers (cali-brated against a laboratory thermometer). O n each visit to the sampling loca-tions I recorded the temperature at the time of observation and the maximum and minimum temperatures for the period since the previous visit. R E S U L T S Direct litterfall The input of all detrital materials was highly seasonal and occurred primarily i n the autumn (Figure 1). During the two months from mid-September to mid-November, more than half the annual l i t terfal l was collected for each stream (56% for Mayfly and Spring Creeks and 63% for Blaney Creek). In general there was little direct litterfall from December through A p r i l for any of the streams and rates were low during late spring and summer. In July and August red alder trees dropped moderate numbers of small, green, and apparently intact leaves i n all years from 1983-1988 (personal observation). The total amounts of litter input were similar between years within each stream. The input to Spring Creek was highest followed by Mayfly and Blaney Creeks (Table II, Figure 1). The total fall of detritus into Spring Creek was more than double that of Blaney Creek and about 1.8 X that of Mayfly Creek. The input of deciduous leaves was similar among streams with annual inputs of 132 - 193 g • 13 m ~ 2 • y r _ 1 (Table II). Mayfly Creek had the highest relative input of deciduous leaves of the 3 streams, accounting for 70% and 74% of the total input for the 2 years respectively. Red alder was the most abundant source of deciduous inputs, accounting for over 69% of deciduous leaf A F D M collected i n the litter traps. The main difference between the 3 streams was i n the amount of input from coniferous species. Spring Creek received 6.9 times the coniferous input of Mayf ly Creek and 16.7x that of Blaney Creek (Table II). In all three streams the input of deciduous leaves was highly pulsed (Figure 2), most entering from mid-September to the end of November. The t iming of conifer leaf fall was also primarily autumnal, especially for red cedar (Figure 3). Hemlock and Douglas F i r leaves (which were not separated, but were mostly hemlock) fell slightly more often during the autumn but continued to drop leaves throughout the year. Benthic Detritus Woody material > 5.0 m m diameter made up the largest portion of C P O M standing crop i n Mayfly Creek during 1985-1986 (Figure 4). Particles > 5 m m from deciduous leaves were much less abundant than wood. C P O M i n the particle size of 1.0 to 5.0 m m was also abundant, but the nature of these particles was not assessed (although the representation i n each category did not appear to be proportional to the amounts of wood, deciduous leaves, or conifer leaves > 5.0 mm). The standing crop of deciduous leaf material showed the seasonal pattern an-ticipated, with higher densities during late autumn and winter than the remainder of the year (Figure 4). There was a 63-fold difference from the peak during January to the minimum of A p r i l . Abundance of leaf particles during summer was slightly higher than spring values. Standing crops of wood (^ 5 mm) and total C P O M were variable over the year, and exhibited seasonal patterns with low summer abundance and higher density during the remainder of the year. 14 5n 4 11 0 fin Mayfly Creek Total - i — i — i — i — i — i — i — i — r --o CM * V) re E 60 # o 4 Spring Creek Total T — I — i — i — i i i i — r — i — s — I — i — i — i — i — i — i -5 4^  3 I-\-0 —I—I—I—T 1984 Blaney Creek Total T 1—I 1 1—I—I— 1985 • i i i 1986 Figure 1 — Daily litterfall rates (g ash-free dry mass - m - 2 - d - 1 ) for all detritus collected i n litter traps along the margins of U B C Research Forest streams. Note the different scales on each ordinate. The rates are averages for each sampling interval. 15 T a b l e I I — Ash-free dry mass of components of direct litterfall to small coastal streams i n g • m - 2 • y r _ 1 . Grand totals at bottom represent annual (July to July) litterfall of all plant detritus. Detritus year 1985 was from 31 July 1984 to 30 July 1985, and year 1985 covered the period from 31 July 1985 to 30 July 1986. Component Mayfly Creek Spring Creek Blaney Creek 1984 1985 1984 1985 1984 1985 Alder 121.6 136.9 122.7 109.3 131.1 123.6 Vine Maple 35.7 30.7 27.5 27.0 3.8 1.2 Rubus spp. 15.8 17.3 18.2 11.6 11.5 3.4 Wi l low 1.7 3.5 0.4 0.1 29.0 10.6 Total Deciduous 1 174.8 193.0 172.2 132.4 150.1 153.4 Cedar 7.2 32.3 114.6 137.3 0.2 0.6 Hemlock, and Douglas F i r 2.5 1.4 49.0 65.2 2.5 18.7 Total Coniferous 9.7 33.7 163.6 202.5 2.7 19.3 Wood 48.1 16.5 32.6 48.8 36.9 10.6 Totals 2 248.8 261.7 448.5 481.5 219.7 201.8 1. Total deciduous leaves includes leaves of species not listed explicitly 2. Grand totals include miscellaneous plant parts 16 Blaney Creek Deciduous "I v ' i 1 | l l i 1985 1986 1984 Figure 2 — Daily litterfall rates (g ash-free dry mass - m " - 2 • d _ 1 ) for deciduous leaves collected i n litter traps along the margins of U B C Research Forest streams. Shading indicates the contribution of red alder. Note the different scales on each ordinate. The rates are averages for each sampling interval. 17 1.0-0.8-0.6-0.4-1 0.2-0-Mayfly Creek Coniferous I/) ro E T3 50 3 0 Spring Creek Coniferous - i — i — i — i — i — i — i — r — i — i — i — i — i — r 0.20-1 0.16 0.12 0.08-0.04-Blaney Creek Coniferous O M P H , I •,, r r r •. •"•. i • 1984 1985 1986 Figure 3 — Daily litterfall rates (g ash-free dry mass - m - 2 • d - 1 ) for coniferous leaves collected i n litter traps along the margins of U B C Research Forest streams. Note the different scales on each ordinate. The rates are averages for each sampling interval. 18 CM A. Deciduous Leaf >5.0 mm 12 E V V i _c IA eo ex. o bO c 5^ c ro to 40 \ B. Wood > 5.0 mm 804 40 C. Total C P O M > 1.0 mm A S O N D J F M A M J J F i g u r e 4 — Estimates from Mayfly Creek of benthic standing crops from August 1985 to July 1986 (g ash-free dry m a s s - m - 2 ± 1 standard error) of a) decid-uous leaf detritus > 5.0 m m diameter, b) wood detritus > 5.0 m m diameter, and c) all detritus > 1.0 m m diameter. For each estimate, sample size was > 5. Note the different scales on ordinates. 19 Detrital materials > 1 m m were heterogeneously distributed spatially which is reflected i n the large standard errors associated with the estimates of abundance. The patchiness of leaf detritus distribution was more obvious during winter, whereas particle sizes were smaller and accumulations smaller during spring and summer. Woody materials were always very patchily distributed considering the large stan-dard errors (Figure 4). Decomposition rates In total 19 separate series of decomposition trials were run i n either Mayfly Creek or Spring Creek (Table III). In all trials a linear model for rate of loss ex-plained a larger proportion of the total variance i n the data set than an exponential model. The nested A N C O V A (Table IV) explained 87.6% of the total variation i n the data set; the same analysis using Zn(dry mass) as the dependent variable had an R2 of 0.816. There was no significant difference i n the rate of decomposition between streams ( P > 0.9) when all other factors were accounted for. The rate of loss of leaf pack material was significantly related to both the number of days i n the stream and the accumulated degree-days (Table I V ) . One potential source of unexplained variance i n the A N C O V A was between season differences after the temperature effect was removed. To assess this I used the residuals from the A N C O V A i n a one-way A N O V A [General Linear Models -SAS (1985)] with season as the independent variable. There was no significant effect of season ( F 3 )386 = 0.06, P > 0.9). To determine the temperature effect on decay rate, the rate of leaf loss per day for each trial (the slope of the linear regression) was regressed against the average temperature during the incubation period. There was a significant linear increase i n the rate of decomposition as temperature Increased (Figure 5, R2 = 0.46, P < 0.002). The slope of the relationship of dry weight remaining versus time 20 T a b l e I I I — Summary of the streams and dates of the alder leaf pack decompo-sition trials. The init ial number of leaf packs are indicated i n brackets after the number recovered. The slopes of the linear and logarithmic (loge) regressions are given, although the slopes of the linear regressions are used i n Figure 5. Trial Stream Dates N Linear Slope Exponent Average Temperature 1 Spring 3 June - 3 Aug 1985 25 (25) -0.055 -0.027 13.2 2 Mayfly 3 June - 3 Aug 1985 22 (25) -0.067 -0.040 12.3 3 Spring 3 July - 3 Sept 1985 25 (25) -0.054 -0.023 13.7 4 Mayfly 3 July - 3 Sept 1985 23 (25) -0.050 -0.023 13.0 5 Spring 5 Sept - 14 Nov 1985 21 (25) -0.023 -0.006 8.7 6 Mayfly 5 Sept - 15 Nov 1985 20 (25) -0.024 -0.007 6.8 7 Spring 5 Oct - 14 Nov 1985 15 (15) -0.016 -0.004 7.4 8 Mayfly 5 Oct - 15 Nov 1985 15 (15) -0.015 -0.004 4.9 9 Mayfly 1 May - 26 June 1986 14 (15) -0.034 -0.011 8.8 10 Spring 1 May - 26 June 1986 15 (15) -0.038 -0.011 10.6 11 Mayfly 15 May - 10 July 1986 15 (15) -0.035 -0.010 9.8 12 Mayfly 29 May - 24 July 1986 15 (15) -0.039 -0.011 10.7 13 Mayfly 12 June - 7 Aug 1986 15 (15) -0.049 -0.019 11.2 14 Mayfly 26 June - 21 Aug 1986 14 (15) -0.028 -0.010 11.8 15 Mayfly 10 July - 4 Sept 1986 15 (15) -0.060 -0.026 12.4 16 Mayfly 7 Aug - 2 Oct 1986 15 (15) -0.064 -0.027 12.6 17 Mayfly 18 Sept - 14 Nov 1986 15 (15) ' -0.038 -0.013 7.1 18 Mayfly 14 Nov 1986 - 24 Jan 1987 18 (18) -0.035 -0.012 2.0 19 Mayfly 27 May - 16 July 1987 57 (60) -0.071 -0.047 11.2 21 T a b l e I V — Results of nested A N C O V A [using partial (Type III) sums of squares from General Linear Models (SAS 1985)] of decomposition rates of red alder leafpacks. Source d.f. M.S. F P Stream 1 0.0011 0.01 0.938 Tr ia l (Stream) 17 0.2704 1.41 0.130 Days x Tria l (Stream) 19 0.7799 4.06 < 0.0001 Degree-Days x Tria l (Stream) 19 0.5962 3.11 < 0.0001 Error 333 0.1920 apparently increases as temperature increases since an exponential curve fit trie data equally well (R2 = 0.46, P < 0.002). The rate of decay of alder leaf packs was modelled with a multiple regression with number of days i n the stream and the accumulated degree-days as independent variables. The number of days did not contribute significantly to the fit of the model (P > 0.1) and only explained 0.1% of the total variance. Degree-days alone accounted for 67.5% of the variance i n dry mass of leaf remaining. The rate of decomposition of 5 g alder leaf packs was dry mass — 4.37 — 0.0041 degree — days (-2*1,389 = 811, P < 0.0001). The intercept was less than 5.0 g due to leaching and handling losses. D I S C U S S I O N The highly seasonal t iming of litterfall into these montane streams is a large contributor to the seasonal variation i n the abundance of C P O M available as food to detritivorous invertebrates. The patterns of high autumnal input rates and abun-dant benthic detritus during winter are generally found i n small temperate streams (Neaves 1978, Webster et al 1983, Connors and Naiman 1984). 22 0.08 n 43 0.06 bO j§ 0.04 % 0.02 O Mayfly Creek • Spring Creek O 0 0 3 6 9 12 Mean Temperature (°C) 15 Figure 5 — Relationship between rate of loss per day (g dry mass-d - 1) for 5 g leaf packs of red alder versus the average temperature during the incubation period. 23 There were large differences between litterfall estimates of input and actual ben-thic biomass of detritus. Deciduous leaves accounted for 72% of annual litterfall to Mayfly Creek, however they represented only 8% of the mean annual benthic stand-ing crop of C P O M . O n the other hand, wood only made up 12.7% of the measured litterfall whereas it comprised 46% of C P O M i n the stream bottom. M y estimates of wood input are surely underestimates since most wood enters as branches and boles > 15 cm diameter so that small litter traps are inappropriate to the spatial and temporal scales of wood input. In addition to this sampling problem, wood decomposes at an exceedingly low rate i n comparison with leaves (Anderson et al. 1978, Anderson and Sedell 1979, Webster and Benfield 1986), so that there is a large amount of between year storage, and fragments from large bole wood can be generated for many years (Triska et al. 1982, W a r d and Aumen 1986). Estimates of benthic detritus from other small, coniferous forest streams indicate > 90% of standing crop is wood (Naiman and Sedell 1979, Triska and Cromack 1980). The benthic standing crop of detritus only roughly reflects the patterns of litter inputs. The rates of decomposition for different materials, as well as the temperature dependence of breakdown rates, result i n seasonal differences i n loss rates. In addition, the timing and magnitude of variation i n the flow regime affect retention of detrital materials within a stream reach (Triska et al. 1982, Cummins et al. 1983). For example, although input rates of deciduous leaves were low during summer, leaf detritus appeared to remain where it fell due to the continuously low discharge i n Mayfly Creek during that period, thus accounting for the increases i n standing crop during summer. Similar deviations between the input and standing crops can be seen i n the data of Short and W a r d (1981) and Webster et al. (1983). Estimates of the loss of P O M due to export (as opposed to decomposition) are scarce, however, one such estimate (Cummins et al. 1983) from published data is 24 that 2 - 9 % may be exported as P O M . The loss due to export from the coastal, montane streams is likely to be higher judging from the low standing crops. Leaf material represents a small and variable portion of benthic C P O M , but it is stil l most important as a food source (Cummins et at. 1989). The estimates of standing crops of leaf detritus i n Mayfly Creek were low i n comparison to values from other small streams. Few species of stream macroinvertebrates derive a large portion of their nutrit ion from decaying wood (Anderson et al. 1978). Decomposition rates of red alder leaves increased as temperature increased. This temperature dependence has been suggested previously (Peterson and C u m -mins 1974), but only recently has it been quantified (Hanson et al. 1984, M c A r t h u r et al. 1988, Short and Smith 1989). The important aspect of temperature depen-dent rates of decay is that C P O M is decomposed faster during the time of year when it is least abundant. There are hypotheses that might account for the unexplained variance of leaf decay rates. First , if there is a seasonal succession of consumer species, then demand may change independently of temperature. Second, some nutrient concentrations vary more than an order of magnitude depending on flow regimes and recent pre-cipitation history and are generally higher during winter and spring (Feller and Kimmins 1979, M . Feller - pers. comm.). Nutrient concentrations are known to affect decomposition rates (Howarth and Fisher 1976, Elwood et al. 1981). T h i r d , generally higher flow rates during winter and spring might affect rates of mechani-cal breakdown (Anderson and Sedell 1979). In my data for leaf pack decomposition rates, once the effects of accumulated degree days and actual days on decay of alder were accounted for, there was no evidence of any other seasonal effect. There were no significant differences i n decomposition rates of leaves i n the two streams where I studied leaf pack processing (once temperature and time were factored out). This result suggests that between stream differences i n processing 25 rates which are frequently observed (e.g., Stout et al. 1985) may be partly a result of temperature differences which are rarely accounted for. The litterfall data indicates annual amounts of input typical of the published range of inputs (Anderson and Sedell 1979, Conners and Naiman 1984). Estimates of litter input for small streams vary from about 200 to almost 800 g • m ~ 2 • y r _ 1 (Connors and Naiman 1984). The values I obtained for annual litterfall compare well with Neave's (1978) estimate of 467 g • m - 2 • y r _ 1 for Carnation Creek, Bri t ish Columbia. I did not measure the contribution of lateral movement (a hillslope process) to allochthonous inputs. Some studies have reported as much as half, although usually somewhat less, the detrital input to small streams i n the form of lateral movement (Triska et al. 1982, Connors and Naiman 1984, Naiman et al. 1987). If lateral movement contributes any amount close to 50% of inputs, then these montane Bri t ish Columbian streams would receive annual detrital input higher than many streams. The different litterfall totals between streams and the relative constancy within streams between years are related to differences i n the canopy. Spring Creek is the only stream with a completely closed, coniferous canopy of western hemlock and red cedar. The riparian forest along Spring Creek is about 250+ years old and the major difference i n litterfall rates between Spring Creek and the other two streams was i n the amount of coniferous leaf input. While Mayfly Creek is surrounded by coniferous forest of 60 years of age, this alluvial stretch has active banks with dense growths of deciduous trees such as alder and vine maple on the disturbed patches. The young age of the coniferous trees surrounding Blaney Creek account for the low rate of conifer needle inputs to this stream. Whi le input rates of deciduous leaf detritus to montane streams were low dur-ing summer, there was some direct litterfall of green leaves of alder. Entry of green leaves to streams has been referred to as the result of beaver activity (Stout et al. 26 1985) or severe storms (McArthur et al 1986). Red alders appear to drop small numbers of leaves during July, August, and September i n Bri t ish Columbia. This pattern of summer leaf drop was observed each year from 1983 - 1988 but to my knowledge has not been noted i n the literature on streams. However, it apparently occurs elsewhere, as it has been observed i n streams of the Cascade Mountains of Oregon (Gary Lamberti - pers. comm.). The addition of small amounts of leaf ma-terial during summer is likely to be important to the energetic subsidy of detritivore populations, given evidence that these animals are food l imited (Chapter 4). I have not considered qualitative changes i n detritus over time. Qualitative differences within and between detrital types are known to lead to large differences i n consumer growth rates (Otto 1974, Sweeney and Vannote 1986). However, these qualitative aspects are difficult to quantify. After an ini t ia l conditioning period when quality per mass increases because of microbial colonization (Kaushik and Hynes 1971), the remainder becomes increasingly composed of highly refractory compounds such as lignin (e.g., Triska et al. 1975, Barlocher 1983). Thus, changes i n resource quality are difficult to translate into resource availability to consumers i n natural systems. These results demonstrate some of the patterns of litter entry and decom-position that contribute to seasonal changes i n the abundance of coarse detrital materials. There are many other aspects to the dynamics of detritus, such as qual-ity changes, burial (Cummins et al 1983), and heterogeneity of distribution within and between reach types (Lamberti et al, in press). The three primary sources of seasonal variation i n detrital standing crop are timing of l itterfall , temperature dependence of decomposition rates, and winter and spring flushing associated with increased discharge. The variation i n abundance bf assimilable detritus sets the pattern of resource availability to detritivores. Some of the consequences of this 27 seasonal variation of detritus standing crops for detritivore populations w i l l be ad-dressed i n chapter four. 28 C h a p t e r T h r e e NATURAL HISTORY OF STREAM DETRITIVORES Details of the biology of most species of stream invertebrates are poorly known, i n part because immature stages last for upwards of 90% of the life cycle for many species. In addition, immature stages are often not identifiable to species. For most of the species I have investigated there were no detailed reports on their natural history except for a few records of emergence dates and geographic distribution. In this chapter I present data on life cycle patterns and other observations of the detritivores collected i n Mayfly Creek, Bri t ish Columbia. One goal of examining life cycle t iming is to consider how phenology corresponds to seasonal changes i n food abundance and the timing of life cycles of other species. M A T E R I A L S A N D M E T H O D S The data i n this chapter were assembled from a number of studies, using a vari-ety of sampling techniques. Sources of benthic samples included a regular sampling program of Mayfly Creek i n the University of Br i t ish Columbia Knapp Research Forest near Haney, Bri t ish Columbia. Samples were usually collected monthly us-ing a metal cylinder equipped with staged nets of 250 /xm and 102 fim mesh size, and enclosing a sample area of 712 c m 2 . Data from benthic samples of the Mayfly Creek experimental streams were also included. Data for rarer species also came from leaf pack decomposition and colonization experiments, and from miscellaneous collections of adults. Data on the timing of adult emergence came from a series of emergence cages on experimental stream channels on the floodplain adjacent to Mayfly Creek (de-scription i n Chapter 4). These adults came from a food supplement experiment, but since there were no differences i n phenology due to experimental treatment (Chap-ter 4), I expected the data to be representative of the population i n the natural 29 stream. The emergence cages were emptied at least once every three days during the period M a y to September 1986, and at less frequent intervals thereafter unt i l 30 A p r i l 1987. Individuals i n samples were sorted from debris, identified to species when pos-sible (exceptions are noted i n the Results section), and the width across the head at the level of the eyes (interocular distance) measured to the nearest 0.024 m m with an eyepiece micrometer i n a dissecting microscope. R E S U L T S Some of the details of the life cycle phenology of the most common detritivores i n Mayfly Creek are summarized i n Table V . These nine taxa represent five families in. three orders. The emergence periods of all taxa combined include most of the year. Two groupings can be discerned, those which emerge during winter and early spring (the "winter stoneflies" Zapada spp., Soyedina producta, and Capnia spp.) and those which emerge during late spring, summer and early autumn. The most abundant non-chironomid detritivore i n Mayfly Creek was Zapada cinctipes . The life cycle of this species is clearly univoltine i n this stream (Fig-ure 6). Adul t emergence occurred from 8 January to 7 A p r i l , wi th about half of the individuals emerging during the first half of February (Figure 6). Males appeared to emerge before females, and males were on average 42% lighter than females. The early instar larvae appeared during M a y and June. Zapada haysi is also univoltine i n Mayfly Creek. The adult emergence period occurred between 17 March and 27 May, with most adults emerging 7 - 2 7 A p r i l (Figure 6). Early instar larvae began to appear during the latter half of June, and larvae were slightly smaller than the congeneric Z. cinctipes throughout its life cycle, although at emergence Z. haysi was actually larger and heavier. The Capnia were evidently represented by two species based on two discrete sizes of females collected i n emergence cages. I was not able to distinguish larvae, 30 T a b l e V — Summary of life cycles of common detritivores i n Mayf ly Creek, B C . Major growth period is the interval during which the average individual increases from 10% of the final mass, to the time of emergence. Taxon Voltinism Adult Emergence Period Major Larval Growth Period PLECOPTERA Nemouridae Zapada cinctipes (Banks) Z. haysi (Ricker) Malenka cornuta (Claassen) M. californica (Claassen) Soyedina producta (Claassen) Capniidae Capnia spp. (including C. nana Claassen) Leuctridae Despaxia augusta (Banks) TRICHOPTERA Lepidostomatidae Lepidostoma roafi (Milne) DIPTERA Chironomidae Brillia retifinis Saether 0.5 early Jan. - early Apr. early Oct. - January mid March - late May Nov. - Feb. late May - early Sept. mid June - early Sept. mid March - early May January - late March mid Aug. - late Sept. June - July f X Nov. - Jan. March - Aug. mid July - early Sept. late May - late July > 2 late June - November? May - October f based on larval data for both Malenka species which, were biased toward the more numerous M. californica. t Soyedina producta larvae were rarely collected, and only during winter and early spring. 31 Capnia spp. 0-'—i i i — | — i — i — i i i — | — i — i i i i — r - ' • 1984 1985 1986 F i g u r e 6 — Seasonal development patterns of Zapada cinctipes, Zapada haysi and Capnia spp. i n Mayfly Creek, Brit ish Columbia. Circles represent the mean head width, and error bars indicate range (longer bars) and standard deviation (shorter bars). Dashed lines indicate periods of adult emergence. Total numbers from which figures were composed were Z. cinctipes - 5179 individuals, Z. haysi - 889 individuals, Capnia spp. - 1509 individuals. 32 nor collect two species of males. The males that were identified were C. nana and presumably most of the Capnia belonged to that species. The life cycle was simple with emergence from early January to late M a r c h (Figure 6). Malenka californica and M. cornuta were indistinguishable as larvae, except for a slight bimodality i n the frequency distribution of head sizes during late June and July. Ear ly instar larvae appeared from October to late winter, and growth was rapid from A p r i l to adult eclosion (Figure 7). The slight decrease i n the lower part of the size range during early spring may represent the hatching of M. californica larvae, since that species had a later emergence date. Since both species are included (in Figure 7), the range of larval head widths is greater than that expected for a single species. The emergence periods for these species also overlapped. Malenka cornuta emerged from 29 M a y to 5 August (one individual was caught between 12 and 16 September); the peak emergence was from mid-June to mid-July. Malenka californica adults emerged from 12 June to 16 September, wi th most emerging i n the last 3 weeks of August. Despaxia augusta was semivoltine at this site with the generations distinct and overlapping (Figure 7). Adult emergence occurred from 13 August to 30 Septem-ber with most emerging i n the last week of August. Early instar larvae were col-lected from late A p r i l through June, but none were found during autumn or winter. Growth was rapid during summer and early autumn but slowed to approximately zero during the winter. Three species of Lepidostoma were collected from the study streams, and of these L. roafi was the most common. A l l three species were easily distinguished as larvae except for the first instars of L. cascadense and L. roafi which both construct cases of fine sand i n the init ial instar. A l l three species were univoltine i n Mayfly Creek and there was no suggestion of any diapause or resting stage. The larvae of 33 1.5 H Malenka spp. T 1.0 0.5 E £ I 0 ~o ro I > ? i T 1 1 1 1 r Despaxia augusta i 1 1 1 r ro 1.04, T O I 0.5 i Io O T O 6 i 1 i h t i ? i t , T T O f t t l ' i f O or o n 1 1 [ 1 1 1 r 1984 1985 "i 1 1 1 r 1986 Figure 7 — Seasonal development patterns of Mdlenka species a n d Despaxia au-gusta i n M a y f l y Creek, B r i t i s h C o l u m b i a . I n the plot for D. augusta the open a n d closed symbols d is t inguish different cohorts , otherwise s y m b o l s are as i n F i g u r e 6. T o t a l n u m b e r s measured were Malenka spp. - 1528 i n d i v i d u a l s , Despaxia augusta - 3099 34 L. roafibeg&n. to hatch during the late summer and apparently some individuals con-tinued to join the population unt i l January (Figure 8). The pattern of development was straight forward so the frequency distribution of instars were pooled for many dates i n Figure 8. Lepidostoma roafi adults emerged from 24 June to 6 September, and most left the stream between the last week of July to mid-August . Table VI — Size (mm) and mass (mg) equations for each taxon for use i n calcu-lating biomass from head width measurements. Species Relationship Zapada cinctipes mass = 0 .008e 3 - 1 9 5 ^ e a d Zapada haysi mass = 0 . 0 1 3 3 e 2 " 6 / t e a d Malenka spp. mass = 0 .0047e 3 - 8 5 7 ? i e a d Despaxia augusta mass = 0.0052e 4- 1 6 7 , i e a< i Capnia spp. mass = 0.0039e 4 - 5 4 9 f e e a ( i Lepidostoma roafi mass = 0 . 0 0 6 7 e 5 1 2 6 / i e a d I did not sort or identify adult chironomids, so the description of the life cycle of Brillia retifinis is based only on the collections of larvae and pupae (Figure 8). Few individuals were found i n winter and early spring collections, and none of these were first instar larvae. First instar larvae were collected from M a y to November. Pupae were collected from late June to November (Figure 8). The adult emergence period was inferred from the collection dates of pupae. Regressions of mass on head width were developed for the larvae of most of the common detritivores (Table V I ) . These relationships were used to generate estimates of biomass i n the food supplementation experiment (Chapter 4). Zapada cinctipes have been collected i n detailed studies i n a number of loca-tions. I used data on emergence dates from the literature to examine latitudinal and 35 V - IV Lepidostoma roafi II 1 0 0 % J F M A M J J A S O N D IV ro i n — II Brillia retifinis 100% i 1 r I T r J F M A M J J A S O N D F i g u r e 8 — Seasonal development patterns of Lepidostoma roafi and Brillia retifi-nis i n Mayfly Creek, Brit ish Columbia. The width of each section of each histogram represents the relative frequency of that instar i n collections for that date. The dashed line on the L. roafi plot indicates periods of adult emergence; the symbols on the abscissa are roman numerals for larval instar number and P for pupae. The dashed line on the B. retifinis plot represents the dates on which pupae were collected. Total numbers measured were L. roafi - 452 individuals, B. retifinis- 7295 individuals. 36 altitudinal patterns i n the timing of emergence of this species (Figure 9). There was no simple relationship between the length of the emergence period and ei-ther geographic variable (multiple regression, F2is = 1.075, P > 0.3). The first date of emergence was significantly related to both variables (multiple regression, i < 2 8 = 5.5, P < 0.05, R2 = 0.58). The first date of emergence was later as either altitude (T = 3.072, P < 0.02) or latitude (T = 2.815, P < 0.025) increased. D I S C U S S I O N The life cycles of species considered i n this study were fairly simple and i n general synchronized within a species. In most species development was direct, i.e., no diapause stages. Presumed hatching dates reflected the variation i n adult emergence dates, suggesting no delayed egg hatching. The exception to this pattern was an apparent egg diapause i n Despaxia augusta. Larvae of Soyedina producta were rarely collected and then only during late autumn and winter, from which one could infer an egg diapause. For the majority of these species there is little or no detailed information with which to compare my results, wi th the exceptions noted below. Zapada cinctipes and Z. haysi are both common detritivores found i n mountain streams of western North America (Baumann et al. 1977, Short et al. 1980, Obern-dorfer et al. 1984). Reports on the life cycle of Z. cinctipes from other locations suggest it is usually univoltine (Ellis 1975, Kerst and Anderson 1975, Cather and Gaufin 1976, Sakaguchi 1978). The emergence period of Z. cinctipes apparently began over two months later i n an Alaskan stream and extended into late July (Ellis 1975), compared with a January to March emergence period i n southwestern Bri t ish Columbia (this study). Differences i n latitude and altitude among the loca-tions of these various populations account for much of the geographic variation i n the timing of emergence. 37 Dec Jan Feb Ma r First Date of Emergence F i g u r e 9 — Patterns of geographic variation in the timing of the first date of emergence of Zapada cinctipes summarized from literature reports. Note that the independent variables are plotted on the Y-axis for presentation. References include: f a t h e r and Gaufin 1976, 2 Cl i f ford 1969, 3 EUis 1975, 4Jewett 1959, 5 Kerst and Anderson 1975, 6 Radford and Hartland-Rowe 1971, 7 Ricker 1943, 8 Sheldon and Jewett 1967, 9 this study, f - point represents two records. 38 Zapada haysi hatched approximately six weeks later than Z. cinctipes on av-erage. During most of the growth period the mean size of Z. cinctipes larvae was slightly larger than Z. haysi, but there was considerable overlap. The food and space resource use of these two species is almost entirely overlapping except for a slight difference i n the timing of adult emergence and egg hatching. Malenka californica is found throughout the Rocky Mountains from northern Br i t i sh Columbia to New Mexico (Jewett 1959). The recorded periods of emer-gence vary among populations but are always late summer and autumn (Ricker 1943, Jewett 1959, Sheldon and Jewett 1967, Kerst and Anderson 1975). The con-generic M. cornuta has a more l imited range i n the Coast and Cascade mountains from Brit ish Columbia to Oregon (Jewett 1959). The known emergence records for M. cornuta are late spring to the end of summer (Ricker 1943, Jewett 1959, Kerst and Anderson 1975). As was found for Zapada, the timing of growth and emergence by the two species was almost entirely overlapping. Adults of Despaxia augusta emerged i n autumn and there was no evidence that adults overwintered before laying eggs. There are no records of collections of adults i n the winter or spring (Ellis 1975, Baumann et al. 1977). This suggests that this species has an egg diapause that lasts through the winter, with hatching taking place with the beginning of spring. Early instars were never collected during the late autumn or winter, so it is unlikely that larvae hatch and remain inactive during this period. Collections of D. augusta adults from a stream i n Alaska were made from the second week of August unti l the first week of September (Ellis 1975), very similar to the observations of emergence from Mayfly Creek. The case-building larvae of Lepidostoma roafi were the most dense of the de-tritivorous caddisflies. There are no detailed studies of the biology of L. roafi, aside from work on biosystematics. In one study i n Oregon, seven males of the species were collected i n emergence traps during late August and September (Anderson 39 and W o l d 1972) somewhat later than at Mayfly Creek. In Mayfly Creek there were three species of Lepidostoma, L. roafi, L. unicolor, and L. cascadense. The latter two species had a similar growth period and emerged about a month before L. roafi. Lepidostoma roafi was quite dense i n local accumulations of coarse detritus i n large pools i n the late spring along with L. unicolor and L. cascadense (see also Winterbourn 1971). The number of generations each year for Brillia retifinis was impossible to dis-cern from the data, i n part due to overlapping generations. It apparently had several cohorts annually and this was evident i n the exponential growth of population den-sities over a three month period i n stream channels where food was experimentally supplemented (Chapter 4). The extended periods over which first instar larvae and pupae were collected suggests several generations. Authors studying detritivorous stream invertebrates frequently make the state-ment that life cycles have " . . . become synchronized to the autumnal input of leaf m a t e r i a l . . . " (Petersen and Cummins 1974). The interpretation of this statement is frequently contradictory (Petersen and Cummins 1974, Hanson et al. 1984) and makes no useful predictions for life history timing. Given that any two periodic phe-nomena with an annual cycle wi l l appear synchronized makes it difficult to generate a non-tautological prediction. One way to assess this claim is to consider the t im-ing of detritivore life cycles relative to the seasonal changes i n detritus abundance. Considering the seasonal differences i n the timing of growth of the common Mayfly Creek detritivores, it is difficult to see any pattern i n life cycle t iming related to seasonality of food abundance. The periods of rapid growth for most species overlapped considerably with other species. The growth periods of all nine taxa together spanned the entire year. The two pairs of congeners, Malenka spp. and Zapada spp., were most similar i n life cycle timing to their congener. Thus, there is no reason to suggest temporal separation of 40 resource use. W h e n one considers that there were many other detritivores of lower density (Appendix 1), although frequently of higher individual mass, there would be considerable demand for food throughout the year. The next chapter focusses on this question of detrital supply and demand. 41 Chapter Four SEASONAL FOOD LIMITATION OF DETRITIVORES IN A MONTANE STREAM: AN EXPERIMENTAL TEST. Resource hmitation is a necessary condition for competition (Ti lman 1982) and a central assumption for many models of territoriality and evolution by natural selection. Given the theoretical importance of this assumption, it is surprising for how few systems the role of resource l imitation is understood. There are several kinds of data that are often interpreted as evidence for food hmitation: positive correlations between densities or productivity of consumers and their food resources, negative correlations among density and average individual biomass, and energetic supply-demand budgets approaching a value of unity (Neill 1975, Eadie and Keast 1982, Roughgarden 1986). Most species live i n environments where resource abundance or productivity varies seasonally and can l imit components of demographic rates (Fretwell 1972). Both the amplitude and frequency of resource variation, as well as life history char-acteristics of the organism, w i l l i n part determine the species' demographic responses (Boyce and Daly 1980, Cushfng 1986). Superimposed on these dynamics are the effects of other species (interspecific competition, predation, etc.) and the physical environment (flooding, temperature, etc.). The interactions of all these variables make the outcome i n terms of population densities and community composition difficult to predict (Yodzis 1988). In most small streams, allochthonous detrital materials from riparian vege-tation, mostly leaf litter, makes up the largest portion of the energy available for stream organisms (Chapter 2). However, most of the research on detritus-detritivore systems has focussed on the rates of mineralization of detritus, energy flow, or growth rates of detritivores on different detrital types. M a n y species of freshwater macroinvertebrates depend on coarse particulate organic matter ( C P O M , detritus 42 > 1.0 m m diameter) for most, if not a l l , of their nutri t ion and energetic demands (reviewed by Anderson and Sedell 1979, Currrmins and K l u g 1979). The standing crops of assimilable C P O M , excluding large pieces of wood, may vary seasonally by up to two orders of magnitude (Minshall 1967, Short and W a r d 1981, Barlocher 1983, Webster et al. 1983) due to several processes (discussed i n Chapter 2). The result of variation i n the availability of detrital food is that species which rely on this food source may be seasonally food l imited. Several lines of evidence suggest that lotic detritivores might be food l imited (Anderson and Sedell 1979, Shiozawa 1983). First , the common assertion that the life cycle t iming of detritivores is "synchronized" with the flush of detrital resources (Petersen and Cummins 1974, Cummins et al. 1989) suggests that resource availability may be an important gradient i n the natural selection of phenology. Second, positive correlations between invertebrate density and the standing crop of C P O M (within and between dates) have been shown a number of times for detritivores (e.g. Egglishaw 1964, Curnmins et al. 1973, Drake 1984). T h i r d , short-term supplements of organic materials have resulted i n aggregative effects leading to higher local densities of detritivores (Warren et al. 1964, Mundie et al. 1983). Last, i n some instances, individuals from aquatic invertebrate populations gained more mass and greater l ip id reserves when fed higher quality food than individuals that fed upon ambient detritus standing stocks (Otto 1974, Sweeney and Vannote 1986, Gee 1988). A l l these lines of evidence suggest food l imitation of stream detritivores. The objectives of this study were to test the hypothesis that populations of stream detritivores are seasonally food limited under natural conditions and to determine the demographic consequences of food l imitation. Under this hypothesis, food supplementation should increase mean individual size and fecundity (measured as mass at emergence), raise densities, and increase colonization (mediated through reduced emigration). To test this hypothesis I experimentally manipulated the input 43 rate of leaf litter to stream channels beside a small montane stream i n southwestern Bri t ish Columbia. M A T E R I A L S A N D M E T H O D S The results reported here are from a food supplementation experiment which began 1 M a y 1986 and terminated 30 A p r i l 1987. A similar experiment was con-ducted i n 1985, but late autumn freeze-up terminated the experiment prematurely; data for benthic densities from the 1985 experiment were similar to those from 1986 and w i l l not be discussed further. The experiment involved 4 treatments, each replicated twice, using 8 experimental stream channels. Six of these experimental streams (each 15 m X 35 cm) were built on the floodplain adjacent to Mayfly Creek. The water was supplied from Mayfly Creek 100 m upstream of the channels and passed through two settling boxes before entering the streams. This arrangement allowed access for the immigration of invertebrates to the experimental streams, without accumulating mineral sediments. The intake housing i n Mayf ly Creek was covered with an aluminum grating with holes 7 m m diameter to prevent the ac-cess of cutthroat trout (Salmo clarki) which occurred rarely i n the natural stream. Streamside experimental channels were underlain by several layers of polyethylene and separated by bags of gravel over which the plastic sheets were placed. Substrate was 3-4 cm rounded gravel placed to a depth of 10 - 20 cm i n each channel, with the interstices mostly filled with sand (< 5 m m diameter). A t the downstream end the water was restricted to a small overflow which created a pool of about 2 m length i n each channel. The average slope of the channels was 0.03m/m. The flow rate through these channels varied seasonally, but was not measurably different among channels; the range of discharge estimates among dates was 1.3-3.0 L • s _ 1 (for the period 15 M a y to 4 September 1986) through each channel. Two further channels, 44 similar i n design to the streamside channels (substrate, slope, width , depth of sub-strate, underlain by plastic) were built into the streambed of Mayf ly Creek (referred to as instream channels). The input rates of whole leaf detritus for three of the treatments are shown as daily rates i n Figure 10. The instream channels had no manipulation. The " H i g h " input rate was equivalent to the maximum autumnal input rate and was 40 X the natural summer input rate. The "Natura l " input rate was chosen to mimic that of natural input rates to Mayfly Creek measured from July 1984 to July 1986 (Chap-ter 2). The "Intermediate" input rate was chosen arbitrarily to provide another level of food supplementation; the summer input rate of this treatment was 9x that of natural rates. The "natural" treatment wi l l be referred to with quotation marks to distinguish it from mention of observations i n Mayf ly Creek. The leaf material for addition to the experimental streams was 90% red alder leaves (Alnus rubra) and 10% vine maple leaves (Acer circinatum) by mass. The choice of alder and vine maple as supplemental food was based on data that deciduous leaves accounted for > 70% of the litterfall into Mayfly Creek; alder and vine maple made up 63% of total input (Chapter 2). Leaves were collected i n the U B C Research Forest during autumn shortly after abscission, dried i n the laboratory, then sealed i n plastic bags and stored i n a deep freeze (-20 C) unt i l needed. Leaves were added every other week, immediately following the regular benthic sampling. O n 17 M a r c h 1986, the experimental streams were flushed using a portable pump to reduce the residual coarse detritus. To stock the stream channels with invertebrates, on 17 A p r i l approximately 1200 Surber samples from Mayfly Creek (equivalent to 27 m 2 of stream bottom) were randomly and equally allocated among the six channels. Colonization via the inflow from Mayfly Creek also resulted i n a more or less natural benthic community (see Results - Initial conditions). The experiment began 1 M a y 1986. Ten benthic invertebrate samples were collected from 45 High CN " 'E V) ro E >> ~o hO <D 03 CH + J 3 2 1 0 "Natural" Intermediate M J J A S 0 N D J F i g u r e 10 — Rates of leaf litter input to stream channels for the three manipula-tive treatments (g dry mass - m - 2 • d _ 1 ) . Values are for alder and vine maple combined (90:10 by mass). 46 each of the six streamside channels, and five samples from each of the two instream channels. After collecting the samples, the streamside channels were randomly assigned one of the three leaf addition treatments, and appropriate amounts of leaf litter were added to each stream. Benthic samples were collected using a small Surber Sampler with a sampling area of 402 c m 2 and a net mesh size of 102 /xm. Substrate delimited by the sampler was disturbed to approximately 10 cm depth. After the start of the experiment, 5 samples per stream channel were collected at 2 week intervals; each set of 5 samples removed animals and detritus from an area equivalent to approximately 5% of the bottom area of each channel. Samples were preserved with formaldehyde i n the field and stored for later sorting. In the laboratory samples were washed through a series of sieves, and animals were separated from detritus under a dissecting microscope at 6x and 12 x magnification i n succession. Only animals retained on sieves > 0.47 m m mesh size were counted. Head widths of animals were measured across the level of the eyes to the nearest 24 / /m with an ocular micrometer and the animals were stored i n ethanol. Not every sample was counted. Representatives of each species (fresh or formaldehyde preserved) were dried and weighed to derive mass versus head width relationships for standing crop estimates. Immigration and emigration rates for each streamside channel were sampled 24 June - 14 November 1986, by placing nets with a 250 fim mesh size over the inflow pipes or below the outflow such that the entire flow could be filtered. Sample periods were 24 or 48 hours at weekly or longer intervals. Immigration and emigration estimates were never made i n the 2 days after benthic sampling and litter addition, to avoid the confounding effects of these disturbances. I sampled adults emerging from the six streamside channels, wi th six emergence traps per channel. Each trap covered a bottom area of 900 c m 2 and had a polyester cloth skirt at the bottom to provide a seal against animals escaping. Insects were 47 collected in a cup fitted with an inverted funnel on top of each trap. Traps were usually emptied every three days from 1 May - 18 September and less frequently during the remainder of the experiment. Adults were identified, measured, dried and weighed. Adults of Chironomidae were too numerous and included too many species to be economically processed, therefore analyses for Brillia retifinis are based on larval data only. Statistics All the statistical analyses presented are for data from the six streamside chan-nels. Analyses were done using the General Linear Models procedure for unbal-anced, nested ANOVA designs in SAS version 6.02 (SAS 1985). After fitting the complete model, the treatment mean square was tested using the between stream mean square as the error term (since the two streams per treatment are the repli-cates) to determine the probability of any treatment effects. This procedure is analogous to a repeated measures ANOVA design. In all cases the partial (type III) sums of squares were used. In all tests the critical level of significance was a = 0.05. Data for adult mass at emergence was first fitted to a multiway, nested analysis of covariance, with treatment and sex as main effects, stream (replicates) nested within treatment, and collection date as the covariate. The only interaction term tested was for sex X treatment, since there are examples in the terrestrial literature where only females show evidence for food hmitation (e.g., Juliano 1986, Guyer 1988). Transformation of mass data to ln(n + 1) was done only when it improved the variance explained by the full ANCOVA. The residuals of the mass of the adults of all species considered here were normally distributed by sex, after fitting the appropriate model, according to the Shapiro-Wilk statistic (P > 0.05). The data for gross immigration and emigration numbers, and benthic densities were transformed as log10(n + l) to reduce heteroscedasticity. In all cases treatment effects were tested using the mean square for streams within treatments as the error 48 term, compared to an F-statistic with 2,3 d.f. The least squared means from the fitted A N O V A model for each channel were used for tests of between treatment differences using Tukey's test with a = 0.05. R E S U L T S Life Cycles Details of the life cycle phenology of the common detritivores i n Mayfly Creek are summarized i n chapter three. Most species had a single generation per year without any obvious diapause. Two exceptions to this pattern were the midge Brillia retifinis (Chironomidae) and the stonefly Despaxia augusta (Leuctridae). Brillia was apparently multivoltine at this site, but the cohorts were not clearly defined by semiweekly sampling. Despaxia had a two year developmental period before emerging as adults and appeared to have an egg diapause of about eight months. These nine species of detritivores fall into two groups: those that emerge during the late spring and summer (late M a y to late September), and those that emerge during the winter and early spring (early January to mid M a y ) . This allows a convenient separation of these species into two groups for comparison since their major growth periods are separated i n time. Benthos Comparisons with natural stream. Estimates of benthic densities i n the stream-side channels receiving "natural" input rates, the two instream channels, and the natural bed of Mayfly Creek were compared to assess the effects of microhabitat modification and the damping of discharge variation through the streamside chan-nels (Figure 11). Unfortunately, the two instream channels were filled i n with fine sediments before the start of the experiment as the result of sediment movement during a spate. Benthic densities of most animals i n the instream channels were 49 so low that these channels make a poor comparison with the streamside channels. Statistical comparison of the two streamside control channels with Mayf ly Creek is inappropriate since there is no replication of the natural stream, so only a qual-itative assessment can be made. The data for Malenka spp. (the nymphs of the two Malenka species were indistinguishable) indicate that densities were similar i n Mayfly Creek and the streamside controls during M a y and June (Figure 11). Densi-ties of Lepidostoma roafi were variable and tended to be generally higher i n Mayfly Creek but were less than two-fold more dense i n the natural stream. The category, 'other detritivores' included species which were i n their later instars at the time, as opposed to the abundant, but small individuals of other more recently hatched taxa. The almost four-fold higher densities i n Mayfly Creek i n part represented large species which occurred more frequently i n slower, marginal areas of the stream. Initial conditions. There were no significant differences among the three treat-ments i n the streamside channels i n the init ial benthic densities of any species (Fig-ure 12, A N O V A , all P > 0.50) or densities of all detritivores combined (^ 2,3 = 0.5, P > 0.6). A comparison of all channels using samples as replicates, showed no significant differences (P > 0.75). For most taxa, stream 2, which had been as-signed to the high food treatment, had the highest init ial densities, while the two intermediate treatment streams had the lowest densities. Responses to food manipulations. Larvae of the midge Brillia retifinis showed the greatest increase i n density of any taxon during the experiment. There were significant differences i n densities among treatments (^2,3 = 75.46, P = 0.0027), with all 3 treatments significantly different from each other (Tukey's test) i n the order high > intermediate > "natural" (Figure 12). Density increased as food input rate increased. Based on data to the end of July only, the exponential growth model for the high food populations, with 1 M a y as to (i in days), was Nt = O^SSe 0 0 7 4 * 1 ' — 1 50 F i g u r e 11 — Benthic densities of detritivores i n Mayfly Creek, the instream channels, and the streamside channels. The standard error bars are the sampling error (N>5 in all cases). 51 (F i,i2 = 200.3, R 2 = 0.94). After July 10 the densities of B. retifinis i n the high food channels reached an asymptote of approximately 275 larvae per sample (Figure 12). The data for Malenka spp. nymphs (Figure 12) showed significantly different densities due to treatment (1*2,3 = 10.99, P < 0.05). Densities of these stone-flies i n the high food treatment were significantly higher than the "natural" input (Tukey's test) and greater than i n the intermediate treatment. The intermediate input channels had slightly higher densities than the "natural" treatment. Densities of the cased caddisfly Lepidostoma roafi showed a pattern of increas-ing density with increasing food input rate over the period m i d - M a y to 10 July (Figure 12), but differences were not significant (Figure 12, i<2,3 = 1.01, P > 0.45). Densities of L. roafi for only the period 29 M a y -10 July were also not significantly different due to treatment (F2,z = 2.32, P > 0.2), reflecting the high variability between replicates. There was no effect of food treatment on the densities of Despaxia augusta (Figure 12; ^2,3 = 0.83, P > 0.5). The differences among densities were slight, with intermediate food streams highest, and the high input channels lowest. In contrast to the above taxa, the winter stoneflies Zapada cinctipes, Z. haysi, and Capnia spp. were i n their early instars during summer when most of the benthic sampling was done. As a result, their population densities were very much higher than the summer-emerging species which had already gone through the high mor-tality rates associated with early instars. Densities of both species of Zapada (Fig-ure 12) were significantly different between treatments (Z. cinctipes ^2,3 = 28.76, P < 0.02, Z. haysi i<2,3 = 19.16, P < 0.02). Zapada cinctipes was significantly more dense i n the high food level channels than the other treatments (Tukey's test), and the intermediate channel densities were higher than the controls. The congeneric Z. haysi was more dense under high food inputs than under the other two treat-ments, but the only significant difference occurred between the high and control 52 300-200-100-• — High A- • Intermediate • - "Natural" Brillia retifinis S * • A E u CM O V E c V a c V CO. 6i 4 2 Malenka spp. A A \ ure 12 — Benthic densities of common detritivores from streamside channels following treatment by additions of whole leaf detritus. Each point represents the mean value for the two replicate channels i n each treatment on each sampling date. 53 Figure 12 — Continued 54 channels (Tukey's test). The intermediate channel densities of Z. haysi were higher than the control channels, but not significantly so. The nymphs of the nemourid Soyedina producta were extremely scarce i n the benthic samples and were only found i n late autumn samples, so no analysis of densities was attempted. The densities of Capnia spp. were significantly different between treatments ( F 2 ] 3 = 9.69, P < 0.05). High food inputs resulted i n significantly higher numbers than the "natural" inputs, while the intermediate inputs produced intermediate densities, the effect was not significantly different from the other treatments. Immigration and emigration Numbers of immigrants to the experimental channels for each detritivore taxon, and for the large, summer-emerging species combined, were tested for treatment ef-fects. None of the taxonomic groupings showed any significant treatment effect ( P > 0.05). Immigration of Malenka nymphs was most dissimilar among treat-ments but was not significant (^2,3 = 6.04, P > 0.08). Since numbers of Malenka entering the "natural" channels tended to be greater than the high input channels, immigration cannot account for the increased densities under the high food treat-ment. Immigration of Lepidostoma roafi also was higher into the control channels, followed by the high, and intermediate treatments, but the differences were not significant (P2,3 = 0.84, P > 0.5). Many of the detritivore species had significantly higher gross emigration rates (number per day) from the controls i n contrast to the other two treatments. Gross emigration was used rather than per capita rates because benthic density estimates were so variable that they would have added another source of error to the esti-mates of emigration. Under food supplementation the per capita emigration rate should decrease; since densities i n the added food channels were higher and gross 55 emigration rates were lower, therefore the use of gross emigration was a conserva-tive test of the prediction. Emigration of Malenka spp. from the "natural" treat-ment streams was significantly higher than from the other four channels (Figure 13, jP 2 ) 3 = 75.7, P < 0.003). Likewise, Lepidostoma roafi had significantly higher emigration from the control channels than from the supplemental food treatments (Figure 13, ^ 2,3 = 16.51, P < 0.025). Both species of Zapada had significantly different rates of emigration due to treatment (Figure 14). For Z. cinctipes, the rates were significantly lower from the high food treatment (^2,3 = 21.7, P < 0.02), whereas Z. haysi left the intermediate treatment at a higher rate than the other treatments (i*2,3 = 17.5, P < 0.03). There were no statistical differences i n the gross emigration rates for Brillia retifinis (Figure 14, see below) or Capnia spp. ( P > 0.2). Densities of Brillia retifinis increased so much under food addition that total emigration also increased. To test whether the proportional change i n emigration rates matched the predicted decrease with increasing food, emigration data for B. retifinis were converted to per capita daily emigration rates (Figure 15). From this analysis the high per capita rates of emigration under "natural" food input rates can be seen relative to food supplemented treatments. The differences between treatments are nearly significant (^2,3 = 8.74, 0.06 > P > 0.05). This marginal lack of significance is attributable to differences between replicate channels. Because immigration rates among channels differed slightly, net colonization is a better way to illustrate the actual magnitude of treatment effects on immigration and emigration rates. Net daily colonization (numbers of immigrants minus num-bers of emigrants) generally declined over the summer (Figure 16). Colonization of the channels with supplemented food was significantly higher than the control chan-nels for Malenka spp. (^2,3 = 12.35, P < 0.05) and marginally significant for Zapada haysi (p2,3 = 8.91, P < 0.06). Most other species exhibited non-significant trends 56 0) +-> ro 01 c c, '+-> ro i_ bp "E LU 'ro Q Malenka spp. / \ 4 2-0J / \ / I / \ \ 3 i 2 I !\ Lepidostoma roafi A- • A-July August F i g u r e 13 — Numbers of emigrant larvae of Malenka spp. and Lepidostoma roafi from the streamside experimental channels (number-channel _ 1 -day _ 1 ) . Each point is the mean of the two replicate channels. Symbols as i n Figure 12. 57 •--"Natural" F i g u r e 14 — Numbers of emigrant larvae of Zapada spp. and Brillia retifinis (number • c h a n n e l - 1 • d a y - 1 ) . Each point is the mean of the two replicate channels. 58 £ 0.05 0.044 c o r o i _ hp 1 0.03 r o "I 0.02 £ 0.014 r o Q Brillia retifinis • - "Na tu ra l " A- • • Intermediate • — High / / Augus t F i g u r e 15 — Daily per capita emigration rates of Brillia retifinis. Each point is the mean from the two replicate channels for each treatment, on each sampling date. 59 to higher colonization rates under the high and intermediate food input treatments. It is important to note that for Malenka spp., Zapada spp. and Brillia retifinis the net colonization rates were actually negative. Thus there were net losses of the numbers of these taxa, and loss rates were slower with increased food. A prediction that arises from several models of populations under resource l imitation is that there are qualitative differences among individuals remaining or emigrating from a local population. To examine variation i n the size of emigrants versus the individuals i n the benthos, I compared head widths of Zapada cinctipes, which was the only species with sufficient numbers of emigrants and a continuous distribution of head widths. I used geometric means since head width was lognor-mally distributed. The data for Zn(head width) were used i n an A N C O V A with treatment and channels nested within treatments as main effects and date as a co-variate (total R 2 = 0.95). There was a significant treatment effect (1*2,3 = 101.65, P < 0.002), i n the order high > intermediate > "natural" (Figure 17). To compare benthos and emigrants, I took the means of the residuals for each channel for ei-ther benthos or emigrants. The head widths of emigrants were significantly smaller than benthos (Figure 17, Wilcoxon signed-rank test, P < 0.01, n=6) and this was consistent between channels. Adul t mass at emergence Adult dry mass was measured as an estimate of fecundity, since direct counts of eggs are difficult and time consuming. The effect of treatment on dry mass of newly emerged adults was tested using the mean square between streams as the error term, compared with an F statistic with 2,3 degrees of freedom (see Material and Methods, Statistics). In most cases the fitted A N C O V A ' s accounted for at least 65% of the total variance i n adult mass. From exploratory analyses I observed that most of the interaction terms were not statistically significant; therefore, the only interaction term tested i n the final analyses was treatment x sex. 60 0 ro a: c o v» fO N 'c _o o fO Q <u Malenka spp. • — H igh A — Intermediate • - - " N a t u r a l " . A. \ / \ / \ / Figure 16 — Daily colonization rates of detritivores (number per channel per day) in the streamside experimental channels (immigration — emigration). 61 • — High Zapada cinctipes Intermediate A • - - " N a t u r a l " F i g u r e 16 — Continued. 62 F i g u r e 1 7 — Head widths (geometric means) of Zapada cinctipes nymphs com-paring individuals collected i n benthic samples versus those collected i n em-igration samples: emigrants (filled symbols) and benthic individuals (open symbols) for each treatment. 63 Summer-emerging' species. Food treatments had a significant effect on adult masses of Malenka cornuta (Table V I I , Figure 18). Both sexes were heavier with increasing input of leaf detritus (Table VIII) , although most of the difference among treatments was attributable to females i n the high food treatment reaching a higher mass relative to the other treatments ( P < 0.05). There was a large treatment effect on the mass (data transformed to ln(x + 1), P = 0.002) of Malenka californica, with the increase i n the predicted direction (Table V I I , Figure 18). The A N O V A accounted for 71.5% of the total variance for mass of the adults of this species. After correction for large seasonal changes ( P < 0.0001) (residuals from the regression of mass versus date were added to the grand mean), males showed no increase i n mass i n the intermediate treatment relative to the control, while high food males were significantly heavier ( P < 0.05). Females were also significantly more massive from the high food input channels than specimens from the other two treatments (Table VI I I , P < 0.01). The caddisfly Lepidostoma roafi had the largest proportional increases i n adult dry mass due to food supplementation (Figure 18, Table V I I , P2,3 = 60.9, P < 0.005). A l l three treatments were significantly different from each other, but the major difference was the large increase i n mass of both sexes under the high food input treatment. After correction for the seasonal decline i n body mass, females gained proportionately more mass than males (Table VIII) , although there was no significant interaction between sex and treatment ( P = 0.48). Despaxia augusta was the only common, summer-emerging detritivore to show no significant effect due to food treatments (Figure 18, Table VII ) . Unlike any of the other species there were significant differences between replicate streams ( P = 0.006). Because of the poor fit of the A N O V A model for this species, the percentage differences due to treatment are shown as the least square means of treatments, controlling for sex and seasonal changes (Table VIII) . 64 Malenka cornuta 0.6\ OA-0.2] 0-ri s rt s s \ \ s s s b l.O-ci I 0.8-< 0.6-0.4 0.2 0-Malenka californica ri "Na tu ra l Intermediate H igh F ood Input Level F i g u r e 18 — Mass at emergence (mg dry mass) for common summer emerging detritivores from the experimental streams caught i n emergence traps. Values for males are shown with cross-hatching, and for females with empty bars. Values are arithmetic means ( ± 1 S.E.) after controlling for seasonal effects. Standard error bars based on all individuals. 65 2.0-1.5-1.0 0.5-1 0 Lepidostoma roafi ft S S \ \ \ s \ S \ S \ S S S Q 0.8-1 < 0.6-i 0.4-1 0.2 0 Despaxia augusta s s s s s s s s s s \ s s s s "Natural" Intermediate High Food Input,Level F i g u r e 18 — Continued. 66 T a b l e V I I — Summary nested A N C O V A tables for mass at emergence of adults from experimental streams. Results using partial sums of squares from a general linear model analysis of variance. Treatment effects were tested using Channel(Treatment) as the error term for significance tests. Source d.f. MS F P Malenka cornuta (not transformed) Total R 2 Treatment 2 0.159 14.14 0.030 Channel(Treatment) 3 0.011 1.36 0.258 Sex 1 0.571 69.10 < 0.0001 Sex X Treatment 2 0.016 1.96 0.145 Date 1 0.0001 0.02 0.893 Error 133 0.0083 Malenka californica (mass ln(x + 1) transformed) Total R 2 Treatment 2 1.582 95.53 0.002 Channel(Treatment) 3 0.017 0.45 0.715 Sex 1 6.171 169.03 < 0.0001 Sex x Treatment 2 0.010 0.27 0.765 Date 1 3.258 89.25 < 0.0001 Error 240 0.037 Lepidostoma roafi (mass ln(x + 1) transformed) Total R 2 Treatment 2 1.035 60.90 0.004 Channel(Treatment) 3 0.017 0.53 0.665 Sex 1 4.782 147.88 < 0.0001 Sex x Treatment 2 0.024 0.74 0.481 Date 1 1.948 60.23 < 0.0001 Error 140 0.032 Despaxia augusta (mass not transformed) Total R 2 Treatment 2 0.062 0.72 0.557 Channel(Treatment) 3 0.086 . 4.47 0.006 Sex 1 0.902 46.84 < 0.0001 Sex x Treatment 2 0.124 6.41 0.002 Date 1 0.001 0.06 0.800 Error 99 0.019 67 T a b l e V I I — Continued. Source d.f- MS F P Zapada cinctipes (mass ln(x + 1) transformed) Total R 2 = 0.706 Treatment 2 0.179 25.35 0.013 Cnannel(Treatment) 3 0.007 0.21 0.890 Sex 1 19.508 577.05 < 0.0001 Sex x Treatment 2 0.015 0.44 0.647 Date 1 3.035 89.79 < 0.0001 Error 313 0.0338 Zapada haysi (mass ln(x + 1) transformed) Total Treatment 2 0.393 26.48 0.012 Channel(Treatment) 3 0.015 0.35 0.789 Sex 1 3.231 76.20 < 0.0001 Sex X Treatment 2 0.023 0.54 0.587 Date 1 1.177 27.75 < 0.0001 Error 74 0.042 Capnia species (mass not transformed) Total Treatment 2 0.296 10.00 0.091 Channel(Treatment) 3 0.030 0.69 0.516 Sex 1 1.136 26.26 < 0.0001 Sex x Treatment 2 1.004 23.20 < 0.0001 Date 1 0.070 1.62 0.218 Error 19 0.043 Soyedina producta (mass not transformed) Total Treatment 2 0.115 47.70 0.021 Channel( Treatment) 3 0.002 0.04 0.961 Sex 1 1.636 26.90 0.0002 Error 13 0.061 68 T a b l e V I I I — Percentage change i n adult dry mass relative to adults from the "natural" treatment, based on least squared means controlling for date of emergence. Values for S. producta are based on least squared means for treatments controlling for sex. Capnia are not included since there were apparently two species. The adults of Brillia retifinis were not counted or weighed. Treatment Species High Intermediate Females Males Females Males Malenka cornuta 31 11 6 7 Malenka californica 40 15 12 0 Lepidostoma roafi 46 29 19 8 Despaxia augusta 16 -14 8 -29 Zapada cinctipes 4 9 5 5 Zapada haysi 7 22' -18 6 Soyedina producta 17 -4 69 Winter-emerging species. The most abundant, non-chironomid detritivore i n this study was Zapada cinctipes. There were significant effects of treatment, sex and date of emergence for Z. cinctipes (Table VII ) . The differences i n adult mass between treatments were very small so the least squared means for treatments (controlling for sex, emergence date, and sex x treatment effects) are shown i n Figure 19. The adults from the high food input channels were significantly heavier than those from the control streams (Table VIII) . The congeneric Zapada haysi also had significant treatment effects (Table VII ) . The differences among treatments are shown as least squared means i n Figure 19. Adults emerging from the high food additions increased mass relative to the "natu-r a l " input streams (Figure 19, Table VIII) , but those from the intermediate channels were lighter than the others. The data for Capnia apparently included two species which I could not distin-guish, although most were probably C. nana. Sex and sex x treatment effects were significant (Table VII) . The treatment effect was not significant by itself, but adults from the "natural" input channels were slightly larger. Only 19 adults of the nemourid stonefly Soyedina producta were collected from the experimental streams. When the complete model was fitted, the effects of date and sex X treatment interaction were not significant and the probability of a treat-ment effect was P = 0.06. Removing the date and sex X treatment terms (Table VII) makes the treatment effect significant with P = 0.021, while reducing the model R 2 from 0.822 to 0.773 (Figure 19, Table VII) . The difference occurred between the high input treatment and the other two which were statistically indistinguisable. Com-paring the least square means by treatment (Table VII I , since there were too few specimens to compare by sex) shows that adults from the high food treatment were heavier, and intermediate treatment adults lighter than animals from the controls. 70 1.2 1.1 1.0 Zapada cinctipes -o < 1.5 -j Zapada haysi co i-H -H bO E 1.3 Q 1.1 1.4 1.2 1.0 Soyedina producta "Natural" Intermediate High F i g u r e 1 9 — Mass at emergence (mg dry mass) for three winter stoneflies, showing least square means for treatment (i.e., controlling for sex, replicate, and date). Standard error bars are for mean square of among stream (replicate) variance. 71 Biomass and Adult Numbers To illustrate the seasonal changes i n benthic biomass of the common detritivores as affected by various rates of detritus input, the mean biomass per sample was plotted over time (Figure 20). Under natural rates of input there was an increase i n biomass associated with hatching of young and growth of all individuals. The two supplemented treatments increased only slowly over the controls during May, but the effect of food addition increased i n magnitude during June through August. Standing crops of most species increased i n response to food additions. Over 88% of the increase i n biomass i n the high input channels relative to the controls for the combined dates of 23 July and 7 August was due to increases i n the numbers of the midge Brillia retifinis. A similar comparison between intermediate and natural input rate channels showed B. retifinis responsible for 61% of the increase. B y the 7 August the biomass i n the high food channels was > 8x that of the "natural" channels. To provide another measure of treatment effects on density, I tested the number of adults collected i n emergence traps per stream versus food input rate. Analysis of the log10 transformed numbers and input rates showed significant regressions for some species and i n all cases the slopes were positive (Figure 21). Of the summer-emerging species Malenka californica and Lepidostoma roafi had significantly higher numbers emerging as food input increased. The slope of the regression line was steepest for M. californica. Analysis of the combined numbers of the 4 summer species had a significantly positive slope (Figure 21). The choice of a value for rate of food supply to winter emerging species was difficult for the "natural" treatment since input rate varied through time. The slopes of the transformed numbers of adult Z. cinctipes and Z. haysi versus food input rate were significantly greater than zero ( P < 0.005) whether the natural 72 F i g u r e 20 — Seasonal changes i n benthic standing crop of all nine detritivore taxa considered. Each point represents the mean of the two replicate channels. 73 100-O Malenka cornuta 0 Malenka californica A Despaxia augusta A Lepidostoma roafi 50-I 04— 75 50-25 0-O Zapada cinctipes 0 Zapada haysi A Capnia spp. I 0 1 2 -2 -1 Leaf Input Rate (g • m" • d" ) F i g u r e 21 — Numbers of adults collected from emergence traps for each stream plotted against mean food input rate for A . slimmer species and B . winter species. Rate of input for the "natural" treatment varied over time and was higher for winter species because of the autumnal increase i n loading rate. 74 input rate was integrated for periods M a y to December, September to December, or October to December (Figure 21). The lines relating total dry mass of adults emerged for each channel were plotted separately as summer and winter-emergers (Figure 22). The slopes of the transformed (logio) lines were not significantly different ( A N C O V A , Fit& = 0.97, P > 0.35). There was an increase i n total emerged biomass per stream (excluding B. retifinis) i n the high food treatments versus the "natural" treatment: 3.58 X higher for summer-emergers and 2.44 X for the winter species (a combined increase of 2.95x). The percentage of these increases attributable to changes i n density were 72.3% for summer and 89.3% during the winter. Since Brillia adults were not counted, I used the July and August mean stand-ing crops of third and fourth instar larvae and pupae. When these values were added to the biomass of summer-emerging adults of all other species, the ratio of the estimate of production increased from 3.58 X to 7.05 X (combined annual 4.7 x ) . This is probably a very conservative measure of the production of Brillia. I did not attempt more complete estimates of production, i n part because the voltinism of B. retifinis was not known, and i n part because I was only interested i n relative differences, not absolute values. Other Species Species of the mayfly genus Paraleptophlebia have been shown to be consumers of coarse detritus i n the laboratory by some authors (Sweeney et ah 1986, Mattingly 1987). A n analysis of benthic densities of P. temporalis and P. debilis combined, showed no significant effect of treatment (p2,3 = 0.97, P > 0.45). Not all adults collected i n the emergence traps were measured and weighed, but analysis for dry mass revealed no significant effects of treatment (i*2,3 = 1-94, P > 0.25). For many insects wing length and head width are correlated with mass, so I also analyzed these two variables since most adults had been measured. The data set for winglengths 75 3 < +-» .O 200H ^E % 150 E o 100] 504 O bummer # Winter O 0-0 -2 -1 Leaf Input Rate (g • m" d" ) F i g u r e 22 — Combined biomass of adults collected from emergence traps for all channels, with summer and winter species plotted separately, versus leaf input rate. Leaf input rates for the "natural" input rate treatment for winter-emergers was higher than summer-emergers because of the autumn increase i n loading rates. 76 of adult Paraleptophlebia temporalis (the more common of the two species) also showed no significant differences due to treatment (1*2,3 = 2.7, P > 0.1), but head width was almost significantly different among treatments (i*2,3 = 6.88, P = 0.076), with "h igh" food animals slightly larger than "natural" , and animals from the intermediate treatment smaller. Guild Structure In order to compare relative changes i n species composition under food sup-plementation, the relative percentages of each taxon i n terms of biomass were cal-culated. Comparisons were made for two periods, late spring (Figure 23A) and mid-summer (Figure 23B). During spring the density of most species increased i n response to increased food input rates (see Figure 12), but Zapada cinctipes, the species with the highest biomass under "natural" input rates, increased propor-tionately less than the other taxa (Figure 23A). Most of the relative change was due to disproportionate increases i n the biomass of L. roafi and B. retifinis under food addition. B y mid-summer populations of several taxa had emerged or partly emerged. The relative contribution of Z. cinctipes to overall biomass decreased with increasing detritus input. The primary beneficiary of increased food availability was the chironomid B. retifinis. Biomass of this species increasingly dominated the total biomass of the detritivore guild as food increased (Figure 23B). D I S C U S S I O N This experiment demonstrates that detritivores i n this montane stream are seasonally l imited by food supply. B y supplementing the food available to these consumers, I was able to increase densities, decrease emigration, and increase mass at emergence for several common species. This is the first experimental evidence that lotic detritivores are l imited by the quantity of food available, although sev-eral studies have shown that differences i n food quality can also affect growth of 77 (/) t n ro E o CO. <u bO ro •M C V o GL > •M JO <U 60 4 40 20 0 A • "Natura l " A Intermediate • High • i Z.c. L.r. Mai B.r. Cap Z.h. D.a. B 60 40 20 4 A A Z.c. B.r. Cap Z.h. Mai Figure 23 — Relative detritivore guild composition by percentage biomass for each treatment. Each point is based on the means of two replicate channels on two consecutive sampling dates. A . 29 M a y and 10 June, B . 23 July and 7 August. Taxonomic abbreviations: Z.c. - Zapada cinctipes, L.r. - Lepidos-toma roafi, Mai - Malenka spp. B.r. - Brillia retifinis, Cap - Capnia spp., Z.h. - Zapada haysi, D.a. - Despaxia augusta. 78 detritivores i n natural populations (Otto 1974, Sweeney and Vannote 1986, Gee 1988). The magnitude of the response to food additions indicates a large potential for increased production by this montane stream community. The average mass of adults captured at emergence was the parameter most sen-sitive to treatment effects. In spite of minimal replication, even slight differences were significant for the winter-emerging nemourids. Because changes i n fecundity may occur proportionately faster than changes i n individual mass for some insects (Gilbert 1984), even the small increases i n mass at emergence for the winter stone-flies may result i n increments i n individual fitnesses. Net colonization of the experimental stream channels was higher when food was supplemented. More Malenka left the control channels than entered them, suggesting that resources were less abundant than demand at natural input rates of detritus. In all three treatments the numbers of Zapada cinctipes and Z. haysi emigrants were greater than immigrants, but the net difference decreased with en-hanced food. Negative net colonization reflects the increasing demand incurred with larger average body mass as a result of growth during the summer. Among the summer-emerging species, emigration from the intermediate and high food in-put treatments was negligible, suggesting that both treatments provided sufficient food. Considering the very low rates of immigration and emigration, and mini-mal replication, it is striking that differences i n colonization were detectable among treatments, since even a moderate amount of unmeasured variability might obscure these effects. These estimates of immigration and emigration have less variation than the benthos estimates, since the former measures integrate patterning over a larger spatial scale, thus reducing the noise associated with small spatial scales (benthos samples). Apparently some mechanism tightly coupled with food abun-dance is responsible for measurable differences i n emigration, but to address the mechanism would require more fine scale experiments. 79 One hypothesis to explain movement patterns of stream invertebrates is that emigration (in this case "drift") functions i n part as a means by which individuals move among patches of resources. This food effect has been modelled and validated experimentally for algal feeding stream invertebrates, where drift rates decrease with increasing food quality of a patch (Kohler 1985). Emigration rates from my experimental streams were tightly linked to food abundance, since emigration rates generally declined with increasing food. If emigration is i n part a function of per capita resource availability, then perhaps once density, or biomass, reaches equilib-r i u m with resource supply, excess individuals would not be successful at colonizing or remaining (Lomnicki 1988). This idea has been expressed as the "excess produc-t ion" hypothesis (Waters 1972) to explain the drift of stream insects and my data also support this hypothesis. One prediction of Lomnicki 's (1988) model is that emigrants should be smaller individuals (with less than average fitness) and this is borne out by the observation that emigrating Zapada cinctipes are consistently smaller than residents. In an experimental fertilization of small stream reaches with sucrose, Warren et ah (1964) also found large reductions i n the per capita drift rates of benthic invertebrates i n response to the augmentation of bacterial produc-tion. The late summer decline i n both immigration and emigration rates i n my experiment may reflect ontogenetic changes i n the propensity to drift or reduced densities due to emergence and other losses. The lack of statistical significance of differences i n benthic density for some species may be attributable to having only two replicates. Benthic data from streams are notoriously variable (e.g., A l l a n 1984). The density responses to leaf fitter additions were all i n the predicted direction and presumably increased repli-cation would have shown more statistically significant density differences among treatments for some of those species. 80 The exponential growth i n the population density of larval Brillia retifinis i n response to high rates of food addition matches expectations based on release from food hmitation. The large density increases (> 10 x ) shown by larval B.retifinis indicate that these midges were able to exploit the additional food more than the other species, likely by virtue of their short generation time. Brillia retifinis may fill a role as a fugitive species i n this system, since it was better able to track its resources through short-term variation (sensu Horn and M a c A r t h u r 1972, Boyce and Daly 1980). The asymptote that the density of B. retifinis reached is suggestive of a carrying capacity (Figure 12). Both per capita emigration rate and absolute numbers of B. retifinis leaving the high-food input channels increased at about the same time as the asymptote i n density was achieved. In one Michigan stream, Brillia flavifrons was found to be the dominant detritivore (by biomass) on experimental leaf packs where populations of other C P O M consumers were scarce (Stout and Taft 1985). This observation might be a result of the large capacity for increase of these chironomids where additional food is provided [although Stout and Taft (1985) assumed B. flavifrons was univoltine]. Biomass increased more than 7-fold i n summer and about 2.5-fold i n winter as a result of food additions. More of the increase i n non-chironomid adult biomass was attributable to increases i n numbers among the winter species, compared with the summer species (89% vs. 72% ). This difference may i n part reflect the timing of the start of the experiment, since species such as Malenka spp. and Lepidostoma roafi had already gone through the first few instars when absolute mortality rates are normally high. Because these species are univoltine there was no possibility of a reproductive response after experimental treatments had begun, which would have led to increased numbers i n the spring or summer. Thus, the scope for an aggregative response (positive colonization) by these large summer-emerging species to increased food would be reduced and higher per capita food availability would 81 result for survivors. The small differences i n mean individual adult biomass between treatments for the winter-emerging species probably occurred because of the longer period of colonization which allowed a numerical response. There are two other reasons for the smaller differences i n standing crop among treatments during autumn and winter. First , differences i n treatment levels were much reduced because the "natural" input rates include a large autumnal pulse. This can be seen clearly where the mean adult mass of winter-emergers from the intermediate and "natural" treatments alter rank. Second, I did not consider large species, [e.g., Umnephilid caddisfly larvae (Appendix 1), which gain mass rapidly during late summer and autumn] because their densities were too low to reasonably estimate. Because of the large mean individual mass of these larvae they might contribute a large portion of the winter benthic biomass. The densities of taxa such as Psychoglypha spp. (Limnephilidae) appeared to be higher i n the experimental channels when food was increased (personal observation). The approximately seven-fold increase i n biomass due to food addition dur-ing summer demonstrates the capacity for increased productivity of this system. The actual increase i n summer input rates of detritus i n the high versus "natu-r a l " treatments was 40 X . What prevented fuller use of the added C P O M ? One potential unmeasured source of increased productivity per unit biomass was i n the turnover rate of Brillia retifinis if generation time was decreased with added food. Treatment-related differences i n mean individual mass of B. retifinis were not mea-sured and this might also increase the differences i n production between treatments if individuals with added food reach a higher mass. Higher generational turnover rate and increased individual mass might increase production, but still would not account for less than a 40-fold increase. There are at least three possible reasons for this discrepancy. First , the proportion of C P O M available may decline with increasing mass of leaf accumulation i n packs if physical access, oxygen supply, etc., 82 are reduced within larger packs. Evidence that the rates of decomposition decrease with increased leaf pack size (Reice 1974) suggests that food may not be as readily available i n larger leaf packs. Second, the ability of the detritivore guild to track large, short-term pulses of food may be restricted by annual reproductive periods and low spring densities. Biomass i n the high food channels did increase exponen-tially, but the starting point was so low initially that low spring densities may have imposed a bottleneck on the rate of response to food additions. Final ly, factors other than food may become l imiting above a seven-fold increase i n biomass. Most of the species considered i n this study showed a significant increase i n at least one demographic parameter i n response to food supplementation. However, the lack of any distinct response to increased detritus input by the stoneflies Despaxia augusta and Capnia spp. is curious despite their consumption of this material (Merritt and Cummins 1984). The results of some of the analyses for both these taxa were i n the direction predicted by the hypothesis, but variation was high and effects were small. There was also no identifiable effect of food additions on demographic characteristics of the mayfly Paraleptophlebia temporalis, i n spite of Matt ingly 's (1987) finding that nymphs of this species grow best on a diet of coarse leaf detritus. While particular population parameters may be affected by food additions i n such a direction favouring population growth, there may not be any effect manifest at the population level (Sibly and Calow 1987). For example, increased densities may lead to reduced growth and fecundity. These kinds of life history compensation underscore the caution needed when extrapolating from effects on one population measure to higher level effects (Sibly and Calow 1987). However, there are several arguments for an effect on the population levef from my experiment. Increased mass at emergence and higher densities after food supplemention, which result i n increased total egg production, demonstrate the potential for larger populations. In 83 addition, the response of Brillia to supplemental food can only be due to a popula-tion level response, since more than one generation passed during the experiment. The winter stoneflies, which gained most of their adult mass during autumn when the large autumnal pulse of food was available, were more dense and grew to larger sizes when food was supplemented. There appeared to be a "head-start" ef-fect such that growth rates under "natural" food input rates were lower during late summer and this difference persisted to emergence (see Figure 17). Even species which have life-cycle timing such that the major growth period coincides with au-tumn leaf fall were food l imited by C P O M abundance prior to the autumn resource pulse. Other data from the literature suggest that coarse detritus feeders may be food l imited i n many streams and that this is not just a phenomenon of oligotrophic montane streams. For example, Otto (1974) found that Potamophylax cingulatus (Limnephilidae) larvae caged in situ and fed alder leaves reached a mass at pupation that was 25% heavier than individuals collected from the free-ranging population. In the same study he found that larvae i n the natural population lost a large portion of their stored lipids i n the months A p r i l to June, another result consistent with seasonal food l imitation. A population of Gammarus pulex i n England studied by Gee (1988) showed seasonal reduction i n l ip id stores parallel with food depletion and loss of lipids was alleviated by providing caged individuals with extra whole-leaf detritus. In a study of leaf pack decomposition i n a stream running through a logged watershed (Webster and Waide 1982), there was no decrease i n numbers of detritivores on leaf packs i n comparison with a control reach, and this was at-tributed to leaf packs providing a scarce resource in the logged section of the stream. Similarly, Benfield and Webster (1985) interpretted high densities of detritivores on leaf packs i n second order streams as " . . . islands of food . . . i n the face of declin-ing natural l itter." Grafius and Anderson (1979) derived a simulation model based 84 on observed population dynamics and growth rates and experimentally determined energetic values for a population of Lepidostoma quercina i n an Oregon stream. To make the model fit the observed data required the assumption that larvae were severely food limited for the first 80 days of larval life (Grafius and Anderson 1979), corresponding to the period prior to autumnal leaf fal l . Alternative hypotheses for the increased densities i n my experiment (aside from food) would be that increased leaf fitter provided additional microhabitat for larvae, reduced interference among larvae, or reduced predation rates. M y observations of consumers i n the field suggest that the numbers per leaf are much higher when leaves occur singly than i n leaf packs. A n experiment using artificial leaf packs showed no evidence that leaf packs serve as a refuge only (see Chapter 6). In the present study the majority of species, which were not C P O M consumers, showed no significant effect of leaf litter additions on densities (Chapter 5). If additional leaf htter created special microhabitats then species other than the detritivores should show treatment related responses. Predaceous invertebrates have been considered to be capable of controlling numbers of detritivores i n streams, based on small cage experiments (Oberndorfer et al. 1984). However, if predators are to have any effect i n regulating numbers, there has to be some measurable aggregative effect or increase i n the predation rate i n the short term and probably a reproductive numerical response i n the longer term. I examined my data on predator densities from the detritus manipulations and found no significant treatment effects (Chapter 5). Correlations of predator and prey densities alone are not sufficient to suggest predators regulate prey abundance. Food quantity and quality Food quality, independent of quantity, can affect growth rates and densities of aquatic invertebrates. Experimental manipulations of detrital quality (leaf type) i n a series of small springs produced differences i n productivity and mean individual 85 mass of adults of the nemourid Soyedina carolinensis, i n spite of similar quantity of input (Sweeney and Vannote 1986). Detritivore growth rates i n the laboratory are known to differ when reared on leaves from different plant species, and preference tests indicate consumers choose leaves i n order of their nutritional value (reviewed by Anderson and Sedell 1979, Cummins and K l u g 1979). Alder is generally considered a high quality food because of its high nitrogen content (Triska et al. 1975). Since it is the dominant source of detrital input to Mayfly Creek this likely does not result i n a bias i n the experiment because of its high food value. Fresh, green leaves are of higher nutritional value than senescent leaves i n some cases (Stout and Taft 1985). The leaves I used for additions to the channels were senescent, so green leaves which drop into Mayfly Creek during summer may have higher nutritional value, however, food quality across treatments i n my experiment was controlled for since all streams received the same quality of material. Additions of organic materials to other west coast streams of Nor th America have had consequences indicative of food limitation. Warren et al. (1964) showed that addition of sucrose to sections of an Oregon stream increased benthic biomass by 4.5 X that of unenriched sections, and reduced per capita drift rates. Enrich-ment favoured increases i n chironomids and oligochaetes, presumably because the sucrose stimulated bacterial growth, and because of the short generation time of these animals. Mundie and co-workers (Mundie et al. 1973, 1983, Will iams et al. 1977) have used a variety of organic materials as supplementary food for benthic invertebrate populations i n other coastal B . C . streams. In these experiments there were increases of 1.7 to 4 times the benthic biomass of the controls when grain, soybean, fish food pellets, or fish faeces were added, but the species responding were primarily chironomids and oligochaetes; these organic materials may result i n a very different microbial assemblage than is normally associated with decaying leaf litter. 86 While the benthic standing stocks of C P O M are fairly high compared to leaf input rate, most of this is wood detritus which few invertebrates can utilize, and contributes little to the overall energy budgets of most streams (Minshall 1967, Webster et al. 1983, Webster and Benfield 1986). Wood is thought to be most important for its role i n slowing the flushing rates of other detrital materials (Triska and Cromack 1980). However, while detrital retention may be high i n streams with large amounts of large woody debris, most of the retained leaf material is likely to be refractory, highly lignified detrital components (Barlocher 1983). What small amount of deciduous leaf litter that remains by spring apparently supports much lower biomass of invertebrates per unit mass (Cummins et al. 1989). If so, this implies that seasonal losses of high quality allochthonous detritus is a general phenomenon of temperate streams. Abiotic conditions Disturbance due to flooding has been considered to be an important influence on benthic insect populations i n streams (Stanford and Ward 1983). Variation i n discharge can result i n redistribution of detritus (Cummins et al. 1980) and consumers. While discharge variation was reduced i n the experimental channels (relative to Mayfly Creek), this is not a concern since coastal streams typically have low, stable flows from late spring unti l early autumn (Feller and Kimmins 1979), and conditions were the same i n all treatments. In addition the largest effects observed i n this study occurred during summer. Another potential effect of changes i n base flow rates i n natural streams is that low summer discharge effectively decreases the wetted perimeter of the stream, resulting i n increased densities; this could exacerbate the severity of food limitation i n the natural streams. Competition, coexistence and community organization This study did not address the competitive process directly, and all that can be concluded is that the potential for competition exists among the detritivores 87 i n Mayf ly Creek. M a n y authors have examined differences among lotic taxa and speculated on how these differences may lead to coexistence on l imit ing resources, without testing for resource l imitation. Because the resource supply rate is fixed outside of the interactive system i n the case of allochthonous detritus, there is an incomplete feedback loop from consumers to the producers. Given the low rate of supply during most of the year, especially when respiratory demands are high during summer, this resource can be depleted by consumption and physical breakdown. This means that partitioning by slight differences i n phenology is unlikely since resource renewal rate is low enough that resource depression wi l l persist for most of the year. Aquatic invertebrates i n general have large potential for increased biomass by virtue of their high egg production, big changes i n mean individual mass, and several species i n a trophic level. These potential sources of increased biomass permit the guild to respond to increased food, as shown by the exponential increase i n biomass under food supplementation. Supplementing the supply of alder and vine maple leaf detritus i n my experiment could lead to release from food limitation i n at least two ways. First , consumers can feed directly on the additional food, but alternately the shift to whole leaf detritus may also have reduced competition for other resource types. During spring and summer, resource supply rates of allochthonous detritus are presumably lower than potential rates of consumption so i t is likely that resource depression can occur. Several models have been elaborated to describe population changes within generations that take place when food supply can be depressed, especially i n a donor-control system (Nicholson 1954, Fretwell 1972, Lomnicki 1988) such as stream detrital systems. In these models, lack of a rapid means of feedback between resource abundance (or productivity) and consumer populations can lead to overexploitation of resources and higher per capita mortality rates. In the case of stream detritivores starvation is not immediate when food intake falls below 88 the incipient l imit ing level, so that individuals can continue to consume resources (leading to greater resource depression) before they die from disease, predation or starvation. The lag time created by this temporal separation of resource dynamics and consumer demographic response should be destabilizing i n this system. It can also lead to the somewhat counterintuitive result that predatory losses may actually increase overall survivorship under certain conditions. This predatory compensation hypothesis is consistent with the net lack of effects of fish predation on benthic populations (e.g., Reice and Edwards 1986). The evidence for food l imitation of a number of coexisting detritivores is prob-lematic when niche differentiation appears to be so l imited. In general, most species of detritivores have the same ordering of preference for detritus from different sources. In field colonization experiments most species exhibit similar patterns of distribution (Short et al. 1980). A l l this argues against any form of resource differentiation between species, so that leaf detritus may be considered a poorly dif-ferentiable resource, even though its quality changes seasonally. The seven species i n my study which show strong evidence for food hmitation overlap broadly i n space and time. Even the winter-emerging species, which carry out most of their growth i n late autumn when food is most abundant, were more massive and maintained higher densities when food was supplemented. There were other less dense species of detritivores which also inhabited Mayfly Creek but which were not collected with high enough frequency to consider i n the experiment. In al l there were > 20 species of detritivores (Appendix 1), most of which might be expected to be food limited. It appears that resource partitioning by differences i n phenology, resource dif-ferentiation, or microhabitat are insufficient mechanisms to account for coexistence on a resource type such as allochthonous detritus. The continual redistribution of C P O M by changes i n discharge wi l l make microhabitat differentiation inappro-priate to a large extent. Since exclusive use or sequestration of the resource is 89 unlikely i n this case, and generations are non-overlapping, models such as the lot-tery model (Chesson 1986) or territoriality wi l l not work. The extreme seasonality of allochthonous detritus abundance may provide an additional dimension to coex-istence problems (Levins 1979), since most of the species involved here reproduce annually and therefore cannot track the large increase i n food available during late autumn and winter, providing opportunities for other species. There were distinct changes i n the guild composition i n response to detri-tus manipulations. The relative biomass of the winter stonefly Zapada cinctipes decreased with increasing food input rate. This species appears to be the domi-nant detritivore (by biomass) i n many western montane streams (Short et al. 1980, Oberndorfer et al. 1984), suggesting competitive dominance if food limitation is widespread. During spring, two species increased their biomass disproportionately, Lepidostoma roafi and Brillia retifinis. By summer, B. retifinis was the only species to show a disproportionate increase i n biomass. The summer increase of B. retifi-nis follows the emergence of most individuals of the summer-emerging species, and further supports the interpretation that this is functionally a fugitive species. Studies of detritus-detritivore systems have generally focussed on either rates of detritus decomposition and the role of detritivores i n that process, or comparisons of detrital quality i n terms of consumer choice and the support of consumer growth. This focus has dominated studies of stream detritivore systems with the exceptions noted above. Researchers studying other ecosystems also appear to be intent on the role of detritivores i n nutrient cycling and energy flow, rather than on the population dynamics of detritivores, e.g., terrestrial (Dickinson and Pugh 1974, Swift et al. 1979, Hunt et al. 1987), and lake ecosystems (Webster and Benfield 1986). Studies of marine, soft-bottom detritivores have repeatedly shown these fine-particle feeding, short-lived species to be food limited (e.g., Findlay 1982, Alongi and Tenore 1985, Gee et al. 1985). 90 It was not practical to continue the food supplementation experiment beyond the single year, so it is likely that some of the effects are transient and one would expect to see a different community configuration after a few generations (Yodzis 1988). O n a larger spatial and temporal scale, populations of other species would probably also change, e.g., fine particle feeders might benefit from increased produc-tion of fine detritus via C P O M detritivore activity. If the experiment had continued for several years, other, slower-responding feedback loops may have been expected to continue to alter the community. Nevertheless, the results show that food sup-ply, especially during the summer period, does l imit productivity at the C P O M consumer trophic level. These results show that a community of lotic detritivores can be food limited under seasonal conditions. The species considered here are all common i n most coastal montane streams and appear to coexist despite the potential for strong exploitative competition. There are several important questions that remain to be addressed. First , the mechanisms of competition and the interactions of these detritivores i n a food limited environment deserve exploration. Second, how do other processes such as parasitism or predation affect the competitive interactions? Final ly, what are the mechanisms contributing to coexistence and stability of these species populations? 91 C h a p t e r F i v e INDIRECT EFFECTS OF DETRITUS MANIPULATIONS The importance of indirect effects, i.e., interactions between species mediated through intervening species, has been emphasized i n recent years (Bender et al. 1984, Davidson et al. 1984, Yodzis 1988). There are many possible indirect effects, such as apparent competition (Holt and Kotler 1987), synergistic effects between predators (Soluk and Collins 1988), and exploitative competition. Tests of these effects require detailed manipulations of particular species (Carpenter 1988). The best known hypothesis of an indirect interaction i n streams is a facilitation hypothesis based on detritivorous consumers of coarse particulate organic matter ( C P O M ) making finer particles of detritus ( F P O M ) available to fine particle feeders (Cummins 1974, Short and Masl in 1977, Vannote et al. 1980). Fine particles of de-tritus are generated from feeding activities and defecation of detritivores, physical breakdown of larger particles (Cummins et al. 1983), and by flocculation of dissolved organic carbon leached from detritus (Lush and Hynes 1973). The flow of energy through detritivores to collectors (fine particle consumers - after Cummins 1973, see Chapter 1) is a central assumption of current views of stream ecosystem function (Cummins 1974, Vannote et al. 1980). However, there is little evidence for the as-sertion that generation of F P O M due to detritivore feeding activities contributes to nutrition of collectors under natural conditions (Winterbourn et al. 1981). Several studies have shown that manipulations of detritivore abundance can affect concen-trations of suspended and benthic F P O M (Wallace et al. 1982, Mulholland et al. 1985), but these studies did not look for an effect on collectors. Laboratory studies have shown that F P O M generated from shredder faeces is often better energeti-cally and nutritionally for collectors than the average quality of F P O M sampled 92 from natural streams [Short and Masl in 1977, Ward and Cummins 1979, Mattingly 1987, but see Shepard and Minshall (1984) for an exception], A second possible effect of coarse detritus manipulation is that increased pro-ductivity at one trophic level may result i n increased available prey biomass for their predators. These kinds of "bottom-up" mechanisms have been discussed at length i n terms of lake communities i n recent years (Carpenter et al. 1985, McQueen et al. 1986, Northcote 1988), but have figured little i n stream literature (Bowlby and Roff 1986), although it is an implicit assumption of stream ecosystem theory. Predators are predicted to exhibit aggregative increases i n density within patches of high prey density. If prey availability is increased then capture rate per predator might increase i n addition to, or instead of, increases i n predator density. In chapter four, I described how populations of detritivores benefitted from experimental additions of whole leaf detritus. The detrital supplementation experi-ment provided an excellent opportunity to examine the effects of increased detrital input rates on field populations of species that do not directly consume C P O M . There are several predictions following from the hypotheses described i n the pre-vious paragraphs. If collectors of fine organic particles ( F P O M ) benefit from the feeding activities of large particle detritivores, then there should be increased den-sity or growth of collector species. Invertebrate predators should show aggregative responses to patches of high densities of potential prey, and/or increased growth rates. M A T E R I A L S A N D M E T H O D S Details of the experimental design have been given i n chapter four. Briefly, natural benthic communities were allowed to establish i n a series of streamside stream channels receiving water from the adjacent Mayfly Creek and allowing im-migration and emigration from the channels. Three treatments with two replicates each received different input rates of dried alder and vine maple leaves. One pair 93 of channels had an input rate which mimicked the natural input rates recorded for two years i n Mayfly Creek; this is referred to as the control or "natural" treatment to differentiate it from observations i n the natural stream. The high food input rate was equivalent to the peak autumnal input rate but sustained throughout the supplementation period. Intermediate treatment streams received an input rate 9x that of the summer input rate of the "natural" treatment. Ten benthic samples were collected from each channel on 1 M a y 1986, and then the channels were randomly assigned to treatment. A t two week intervals from M a y to the middle of October and less frequently unti l 30 A p r i l 1987, five samples per stream were collected. Benthic samples were collected using a modified Surber sam-pler wi th a sampling area of 402 c m 2 and a net with mesh size of 102 fim.. Samples were preserved immediately i n formaldehyde solution. In the laboratory, samples were sieved into various size fractions and all animals retained on a 471 /xm sieve were separated from detritus under a dissecting microscope at 12 x magnification, identified to the lowest reliable taxonomic level, and counted. O n three dates during the autumn and winter, only animals retained on a 1 m m sieve were counted and so these data were treated separately in the results. Animals were measured to the nearest 0.024 m m and then stored i n ethanol. Immigration and emigration rates were sampled periodically using nets with mesh size 250 ^ m through which the entire water flow of each channel could be filtered. Samples for immigration and emigration were taken at irregular intervals, but never during the 48 h after benthic samples had been taken and detritus added. Emergence traps were emptied about twice a week from M a y to October and less frequently thereafter. Emergent adult insects were collected i n a cup fitted with an inverted funnel placed on top of each trap. The data sets for many species were incomplete for several reasons. Dipteran adults were not treated at a l l , and mayfly subimagos could not be identified to species where there were more than a 94 single species i n a genus, e.g., Baetis. Many other taxa were seldom collected i n emergence traps. Taxa were assigned to a trophic category on the basis of data i n Merrit t and Cummins (1984) and Pennak (1978). M a n y taxa were omitted because of low num-bers. Taxonomic resolution was determined by ease of identification or availability of diagnostic characters. Statistics The data were analyzed using SAS (SAS 1985). Benthic data and immigration and emigration rates were transformed as logio(n + 1) and fitted to a multiway, nested, unbalanced A N C O V A using general linear models (SAS 1985) with date as a covariate, and streams as replicates within treatments. After fitting a model using partial sums of squares [Type III sums of squares, SAS (1985)], treatment effects were tested by dividing treatment mean squares by the mean square of replicate channels within treatments as the appropriate error term, compared with an F of 2,3 degrees of freedom. In all cases the critical level of significance was a — 0.05. The least square means for each stream, determined from fitting the complete A N C O V A , were used i n the a posteriori Tukey's test to assess treatment differences. R E S U L T S Benthic Densities A total of 35 taxonomic groupings of organisms, presumed not to be consumers of C P O M on the basis of Merritt and Cummins (1984) and Pennak (1978), were tested for an effect of treatment on density. I considered the data for two periods separately: i) from the first post-treatment sample date of the experiment (15 M a y 1986) to 4 September inclusive, and ii) 18 September to 8 January 1987. The second of these two periods was less intensively sampled (less frequently and fewer samples per date per channel). Benthic densities of four taxa revealed significant differences 95 due to detritus manipulations for the spring and summer period, and these were all fine particle consumers, Chironomini (Chironomidae), Simuliidae, miscellaneous Trichoptera and Copepoda (Table I X , Figure 24). In all four taxa the high food treatments had the highest densities, and i n three of four the intermediate chan-nels had densities higher than the control. The densities of Chironomini i n the high detrital input channels were significantly higher than the other two treatments (Tukey's test, P < 0.05). Simuliid larvae and Copepoda were significantly higher i n the high versus the control channels (Tukey's test, P < 0.05). The miscellaneous Trichoptera (including mostly early instar lepidostomatids and limnephilids) densi-ties did not differ significantly according to Tukey's test ( P > 0.05), but densities i n the high detritus channels were greater than twice that i n the other treatments. These analyses should properly be subject to Bonferroni's correction and the results of the reanalysis show none of the taxa to differ significantly among treatments. There were no significant differences i n densities of any taxon due to treatment during the autumn period (Table I X ) . Many taxa were not included i n this analysis either due to low densities or because they were small enough to pass through the 1 m m sieve used for three of the four dates i n the autumn period. The effects of detrital input manipulations on benthic densities of individual taxa of non-detritivores were not dramatic, so I analyzed the trends i n density. The densities under high detrital input were greater than control densities for 30 of 36 taxa [Sign Test, Daniel (1978), P < 0.001], and greater than the intermediate treatment for 27 of 35 taxa (including one tie, Sign Test, P < 0.005). Twenty-four of 36 comparisons were higher i n the intermediate channels than the controls, but this difference was not significant (Sign Test, P > 0.05). To examine density trends within trophic levels, comparisons were made for rank. Predator taxa were more dense in the high-detritus input channels versus the intermediate level channels for 11 of 13 taxa (Sign Test, P < 0.03) and higher than 96 Table IX — Results of tests of treatment effects on benthic densities of inver-tebrate taxa for two different periods. Groupings of taxa into trophic groups follows that of Merritt and Cummins (1984) and Pennak (1978). Rank num-bers are 3=high detritus input rate, 2=intermediate rate, l=control . The use of a coarser sieve for autumn samples resulted i n some taxa being excluded. 15 M a y — 4 September 18 September — 7 January T A X O N -^ 2,3 P RANK 2^,3 P RANK P R E D A T O R S Perlidae (Plecoptera) 3.21 0.18 3 > 1 > 2 0.59 0.61 1 > 3 > 2 Perlodidae (Plecoptera) 3.08 0.19 3 > 1 > 2 0.59 0.61 3 > 1 > 2 Chloroperlidae (Plecoptera) 0.47 0.66 1 > 3 > 2 1.23 0.41 1 > 2 > 3 Dmnella spinifera (Ephemeroptera) 1.48 0.36 2 > 1 > 3 — — — Rhyacophzla spp. (Trichoptera) 1.29 0.39 3 > 2 > 1 2.50 0.23 1 > 2 > 3 Polycentropus + Parapsyche (Trichoptera) 0.17 0.85 1 > 3 > 2 0.36 0.73 3>2 = 1 Tanypodinae (Diptera) 1.95 0.29 3 > 2 > 1 2.09 0.27 3 > 1 > 2 Oreogeton sp. (Diptera) 2.86 0.21 3 > 2 > 1 0.62 0.59 1 > 3 > 2 Ceratopogonidae (Diptera) 2.36 0.24 3 > 2 > 1 — — — Hexatoma sp. (Diptera) 0.91 0.49 3 > 2 > 1 1.07 0.45 3 > 2 > 1 Dicranoia sp. (Diptera) 3.33 0.18 3 > 1 > 2 0.71 0.56 1 > 2> 3 Acari 2.32 0.25 3 > 1 > 2 0.05 0.95 1 > 2> 3 Turbellaria (Platyhelminthes) 0.51 0.65 2 > 3 > 1 3.87 0.15 1 > 3 > 2 C O L L E C T O R S Baeiis spp. (Ephemeroptera) 1.05 0.45 3 > 2 > 1 1.84 0.30 1 > 2 > 3 Paraleptophlebia debilis (Ephemeroptera) 0.62 0.60 3 > 2 > 1 0.36 0.73 3 > 2 = 1 P. temporalis (Ephemeroptera) 0.43 0.69 2> 3> 1 6.54 0.08 1 > 3 > 2 Ameletus sp. (Ephemeroptera) 0.26 0.79 3 > 1 > 2 3.82 0.15 3 > 2 > 1 Serratella iibialis (Ephemeroptera) 2.56 0.22 3 > 2 > 1 0.18 0.84 3 > 2 > 1 Epeorus sp. (Ephemeroptera) 2.09 0.27 2 > 3 > 1 0.74 0.55 1 > 2> 3 Tanytarsini (Diptera) 2.36 0.24 3 > 2 > 1 — — — Chironomini (Diptera) 24.39 0.02* 3 > 2 > 1 — — — Corynoneura spp. (Diptera) 4.55 0.12 3 > 2 > 1 — — — miscellaneous Orthocladiinae (Diptera) 7.90 0.07 3 > 2 > 1 — — — Simuliidae (Diptera) 9.69 0.05* 3 > 2 > 1 0.36 0.73 3 > 2= 1 Wormaldia anilla (Trichoptera) 1.66 0.33 3 > 2 > 1 0.14 0.88 3 > 2 > 1 Ecclisomyia conspersa (Trichoptera) 0.04 0.96 3 > 1 > 2 1.95 0.29 1 > 3 > 2 miscellaneous Trichoptera 8.90 0.05* 3 > 1 > 2 5.06 0.11 3 > 1 > 2 Elmidae (Coleoptera) 1.87 0.30 3 > 2 > 1 2.55 0.23 2 > 3 > 1 Oligochaeta 3.44 0.17 3 > 1 > 2 0.33 0.75 2 > 3 > 1 Ostracoda 0.92 0.49 2> 1>3 — — — Copepoda 10.73 0.04* 3 > 2 > 1 — — — Pisidiidae (Pelecypoda) 5.04 0.11 2 > 1 > 3 — — — G R A Z E R S Hydroptilidae (Trichoptera) 4.58 0.12 2 > 3 > 1 — — — Cinygmula sp. (Ephemeroptera) 0.41 0.69 1 > 2 > 3 5.18 0.11 1 > 3 > 2 Cinygma sp. (Ephemeroptera) 1.13 0.43 3 > 1 > 2 2.72 0.21 3 > 2 > 1 Ironodes sp. (Ephemeroptera) 3.53 0.16 3 > 2 > 1 4.10 0.14 1 > 3 > 2 97 May June July August Figure 24 — B e n t h i c densities of four t a x a of collectors w h i c h were s ignif icantly different due to t reatment . E a c h p o i n t represents the m e a n of the two repl i -cate channels . A r e a of each sample was 402 c m 2 . 98 May June July August F i g u r e 24 — Continued. 99 controls for 10 of 13 taxa (Sign Test, 0.1 > P > 0.09). The ranking of densities i n the intermediate and control channels was essentially even, with seven of 13 higher under intermediate detrital addition rates. Overall predator density was higher with increased detrital input (Figure 25) but not significantly so (^2,3 = 1-69, P > 0.3). The rankings of collector densities among treatments were significantly different i n the order high > intermediate > control. Fifteen of 19 taxa were more dense i n high versus intermediate, and intermediate versus control channels (Sign Test, P < 0.02). High-detritus channels had 17 of 19 taxa with higher densities than controls (Sign Test, P < 0.001). To illustrate the overall effects of the three treatments, the total numbers per sample for all collectors were calculated (Figure 25). There was a significant treatment effect (^2,3 = 211, P < 0.001), with all three treatments significantly different (Tukey's test, P < 0.05) i n the order high > intermediate > control. The four taxa listed as grazers showed no trend i n densities related to treat-ments (Table I X ) . There was no significant difference i n the densities of grazers due to treatment (Figure 25, F2,z = 1.98, P > 0.25). To examine whether detrital manipulations resulted i n changes i n species rich-ness, I tested for treatment effects on numbers of taxa per sample (Figure 26). There was an increase i n richness over the summer with hatching of eggs of many species. Although the number of taxa per sample was higher with higher detrital loading, from 21.3 taxa i n the control channels to 24.1 under high detrital input [least squared means from general linear models (SAS 1985)], there was no signifi-cant effect of treatment (^2,3 = 1-69, P > 0.3). Emigration and Immigration Only one taxon, the mayfly Serratella tibialis, out of 29 taxa had a significant difference i n emigration rate related to treatments (Table X ) . The emigration rate 100 30] •-High A- • Intermediate • - Control Predators 20-10-May ' June ' July *~ August F i g u r e 25 — Benthic densities of total collectors, predators and grazers i n ex-perimental stream channels with whole leaf detritus additions. Each point represents the mean of the two replicate channels. Area of each sample was 402 c m 2 . 101 JU CL E ro CO v_ <u Q_ ro X 30 J • _ 25 - H i g h Intermediate Control Species Richness 20 15 \A.. M / V May ' June ' July August F i g u r e 26 — Species richness measured as taxa per sample for the three detritus input rate treatments. Each point represents the mean of the two replicate channels (based on the means of all samples on that date i n each channel). Values include detritivores and some rare taxa not listed i n Table I X . 102 was significantly higher from the controls than from either of the other two treat-ments (Tukey's Test, P < 0.05), and losses from the intermediate treatment were slightly higher than for the high-detritus input channels. A surprising result was that five of 29 taxa had significant differences i n immi-gration rates to the three treatments (Table X ) . The design of the inflow pipes was such that there was a 25 cm free fall from the pipe into the channels, so animals could not "sample" channels. The five taxa with significant differences did not share a common trend i n order of treatment ranking (Table X ) . The five taxa were the mayflies Baetis spp., Paraleptophlebia debilis, Ironodes sp. and Cinygma sp., and larvae of the cranefly Hexatoma sp. The immigration rates of Baetis spp. were sig-nificantly higher into the high detritus channels than the control channels (Tukey's test, P < 0.05). Cinygma immigration rates were significantly higher into the high food channels than the intermediate treatment channels (Tukey's test, P < 0.05). The nymphs of the mayfly P. debilis entered the control channels at a significantly higher rate than the high food input channels (Tukey's test, P < 0.05). Tukey's test did not reveal any significant differences among treatments for Ironodes immi-gration rates (in spite of the significant A N O V A result). Hexatoma larvae entered the control channels at a significantly lower rate than the other two treatments, and the highest immigration occurred in intermediate input channels (Tukey's test, P < 0.05). For four of the five taxa which had significantly different immigration or emigration rates among treatments, the actual numbers were exceedingly small - average daily rates per channel were Paraleptophlebia debilis 1.2, Cinygma sp. 0.09, Ironodes sp. 0.25, Serratella tibialis (emigration) 0.07, and Hexatoma sp. 0.35. Baetis average immigration rates were much higher than the other four taxa at 32.4 larvae per channel per day. To test for trends of emigration and immigration rates for each treatment for all taxa, I used treatment ranks. There were no significant trends in either rates i n 103 T a b l e X — Results of tests of treatment effects on emigration or immigration rates of invertebrate taxa. Groupings of taxa into trophic groups follows that of Merritt and Cummins (1984) and Pennak (1978). Ranks are as described for Table I X . Emigration Immigration TAXON 2^,3 P RANK 2^,3 P RANK PREDATORS Perlidae (Plecoptera) 1.28 0.40 3 > 1 > 2 0.78 0.54 3 > 1 > 2 Perlodidae (Plecoptera) 2.00 0.28 3 > 1 > 2 0.00 0.99 1>2>3 Chloroperlidae (Plecoptera) 0.01 0.99 3 > 1 > 2 0.29 0.77 1 > 3 > 2 Rhyacophila (Trichoptera) 0.99 0.47 1 > 3 > 2 0.97 0.47 3 > 2 > 1 Tanypodinae (Diptera) 1.63 0.33 1 > 2 > 3 1.43 0.37 3 > 1 > 2 Hexatoma sp. (Diptera) 2.66 0.22 3 = 2> 1 38.50 0.01* 2 > 3 > 1 Dicranota sp. (Diptera) 5.63 0.10 3 > 1 > 2 0.05 0.97 1 = 2 > 3 Ceratopogonidae (Diptera) 2.94 0.20 3 > 1 > 2 4.11 0.14 3 > 2 > 1 Turbellaria (Platyhelminthes) 0.14 0.88 1 > 2 > 3 0.66 0.58 2 > 1 > 3 Acari 1.12 0.43 1 > 2 > 3 0.55 0.62 2 > 3 > 1 COLLECTORS Baetis spp. (Ephemeroptera) 0.20 0.83 3>2> 1 11.66 0.04* 3 > 2 > 1 Paraleptophlebia debilis (Ephemeroptera) 1.21 0.41 3 > 1 > 2 16.67 0.02* 1 > 2 > 3 P. temporalis (Ephemeroptera) 0.40 0.70 1 > 3 > 2 0.20 0.83 1 > 3 > 2 Ameletus sp. (Ephemeroptera) 0.82 0.52 1 > 3 > 2 0.05 0.95 2 > 3 > 1 Serratella tibialis (Ephemeroptera) 14.37 0.03* 1>2>3 2.37 0.24 1 > 2 > 3 Epeorus sp. (Ephemeroptera) 1.25 0.40 3 > 1 > 2 0.75 0.54 3 > 1 > 2 Tanytarsini (Diptera) 0.45 0.67 2 > 1 > 3 0.32 0.75 3 > 2 > 1 Chironomini (Diptera) 1.80 0.31 3 > 2 > 1 0.44 0.68 1 > 3 > 2 Corynoneura spp. (Diptera) 0.83 0.52 3 > 2 > 1 0.30 0.76 3 > 1 > 2 miscellaneous Orthocladiinae (Diptera) 1.56 0.34 3>2> 1 1.09 0.44 3>2> 1 Simuliidae (Diptera) 0.82 0.52 2 > 1 > 3 0.01 0.99 3 > 1 > 2 Elmidae (Coleoptera) 0.62 0.60 1 > 2 > 3 0.21 0.82 1 > 2 > 3 Wormaldia anilla (Trichoptera) 1.45 0.36 1 = 2 > 3 0.69 0.57 1 > 3> 2 miscellaneous Trichoptera 1.39 0.37 1 > 2 > 3 0.01 0.99 3>2>1 Oligochaeta 0.16 0.86 2 > 1 > 3 0.63 0.59 3 > 2 > 1 Crustacea 1.00 0.46 2 > 3 > 1 1.81 0.30 3 > 1 > 2 GRAZERS Ironodes sp. (Ephemeroptera) 4.96 0.11 1 > 3 > 2 8.97 0.05* 1 > 3 > 2 Cinygma sp. (Ephemeroptera) 1.30 0.39 2 > 1 > 3 11.56 0.04* 3 > 1 > 2 Cinygmula sp. (Ephemeroptera) 2.88 0.20 3 > 1 > 2 0.30 0.76 1>2 = 3 104 favour of any treatment for all taxa. The only trend i n the rankings was that the emigration rate of predators was higher from controls than intermediate detritus channels (9 of 10 taxa, Sign Test, P < 0.03). Adults For a few species I had sufficient data on adults collected i n emergence cages to test for treatment effects. Very few individuals had been weighed, so I also tested for differences i n head widths and wing lengths which are usually positively corre-lated with mass i n insects. The data for Baetis spp., Cinygmula sp., Cinygma sp., Ameletus sp., Sweltsa fraterna, and S. occidens were fitted to an A N C O V A with treatment, sex and stream-within-treatment as main effects, and time as a covari-ate, after which the treatment effect was tested using stream within treatment as the error term compared with p2,3- In most cases there were significant effects of sex and date ( P < 0.05). None of the three measures of size differed significantly among treatments for any of the six taxa (Table X I ) . Fine Particulate Organic Matter The abundance of detritus per benthic sample i n the particle size range 471 ^ m to 1.0 m m was measured as ash-free dry mass. While collectors may feed on much finer particles as well , I assumed that detrital abundance i n different size categories would be positively correlated. There was no significant difference i n F P O M per sample related to treatment (P2,3 = 0.32, P > 0.7). However, for four dates I had measurements of 0.471 - 1.0 m m particles from the emigration nets. The F P O M exported from the channels was significantly affected by treatment (P2,3 = 11.99, P < 0.04) i n the order high > intermediate > control. The export rate from the high input channels was greater than twice that of the controls from which it differed significantly (Tukey's test, P < 0.05), and export from intermediate channels was between the other treatments. 105 T a b l e X I — Summary of A N C O V A tests of treatment effects on adult insects captured i n emergence traps, controlling for sex and date of emergence. TAXON Mass Head Width Wing Length -^2,3 P n -^2,3 P n 2^,3 P n EPHEMEROPTERA Baetis spp. 0.09 0.92 62 0.23 0.81 430 0.17 0.85 395 Ameletus sp. — - t 9 0.51 0.64 54 2.44 0.23 43 Cinygmula sp. 0.08 0.92 31 4.19 0.14 118 0.42 0.69 106 Cinygma sp. 0.68 0.60 21 2.95 0.20 35 2.47 0.23 32 PLECOPTERA Sweltsa fraterna 2.47 0.23 29 0.04 0.96 30 0.28 0.78 29 Sweltsa occidens 1.30 0.39 19 0.13 0.88 22 1.82 0.30 21 \— Ameletus mass was not analyzed because of small sample size 106 D I S C U S S I O N Fine particle feeders, or collectors, were positively affected by manipulations of whole leaf detritus. The densities of four taxa were significantly higher with i n -creased C P O M input rates, and densities of most other collector taxa were higher under detritus addition. The increased densities for each taxon were not dramatic and neither were there obvious effects on rates of emigration, or on size of adult Baetis spp. or Ameletus sp. The results of coarse detritus manipulations on indi-vidual taxa of collector may best be described as diffuse. The hypothesis that the effect of detrital manipulations was a result of microhabitat alteration cannot be excluded, however, the results of Chapter 6 suggest that such is not the case. The role of detritivores generating F P O M for collectors through their feed-ing activities is a central assumption of most views of stream ecosystem function (Cummins 1974, Vannote et al. 1980). It has even been argued that the major role of detritivores i n lotic ecosystems is the conversion of coarse detritus to F P O M (Webster 1983). However, aside from laboratory demonstrations where the only source of F P O M was from detritivores (Short and Masl in 1977), there have been no field tests of the assumption. Calculations of the quantity of fine particles pro-duced by detritivore feeding activities have suggested that this is a primary source of F P O M . The results of the detritus-addition experiment indicate that even on a moderately small spatial scale (the 15 m length of the channels), increased coarse detritus additions can lead to increased densities of collectors. The four collector taxa shown to have higher densities with added detritus were likely to be responding to changes in food and not microhabitat. The blackflies (Simuliidae) and the midge tribe Chironomini were rarely found on leaf packs. A n alternative possibility is that these taxa" fed directly upon C P O M ; however, Blackflies and Chironomini are probably incapable of consuming coarse detritus directly. The group of miscellaneous caddisfly larvae include species which consume 107 C P O M i n later instars (e.g., Lepidostoma, Psychoglypha and Ecclisocosmoecus). The early instars of the above caddisfly genera probably feed on F P O M i n general, but when appropriate C P O M is available perhaps they consume it directly. In the latter case it is possible that this group was at high densities due to direct food supplements rather than increased F P O M . Direct consumption of C P O M by the other 15 taxa of collectors is an unlikely explanation for the trend to increased density at higher detritus loading rates. The quality of ground-up leaf detritus or detritivore faeces has been shown to be better, i n terms of support of growth of collectors, than F P O M sampled directly from natural streams (Short and Masl in 1977, Ward and Cummins 1979, Mattingly 1987). Naiman (1983) has shown that much of the fine detrital material i n suspen-sion i n streams is highly refractory lignified 'skeletons'. Other observations suggest that F P O M quality (and perhaps quantity) may limit individual growth or popu-lation size of some stream collectors. Positive correlations between seston quality and density of fine particle collectors have been observed several times (Georgian and Wallace 1981, Petersen 1987, Vallett and Stanford 1987). These observations, coupled with the results of the coarse-detritus supplementation experiment, suggest that collectors are sometimes food limited. Low rates of C P O M input to streams at some times of year may result i n constraints on population densities for some collector taxa. The consumption of C P O M by detritivores should have resulted i n higher rates of accumulation of F P O M with increased leaf additions. The lack of such accumu-lations i n my channels suggests that collectors fed on the particles and prevented buildup of standing crops or that F P O M was exported directly after defecation by detritivores. Although F P O M export was higher at higher C P O M loading, the in-crease i n collector densities suggest that these F P O M losses may occur after passing through collector guts. 108 Predatory taxa showed little response to increases i n detrital input rates. There was a trend to increased densities under higher leaf loading, but not of large mag-nitude. In light of increased densities of potential prey due to detritus addition one might predict a large aggregative response on the part of the predators. The absence of a distinct response may be attributable to two possibilities. First , many predatory stream invertebrates are known to exhibit strong interference interactions among themselves (Hildrew and Townsend 1980, Walde and Davies 1984, Peckarsky and Penton 1985). Interference among predators may prevent local increases i n densities i n spite of increased prey abundance. Second, relatively lower densities of predators i n Mayfly Creek (relative to detritivores and collectors) may have limited the rate of entry to the channels and i n turn may have limited the rate at which predator density could respond. These results suggest that predators probably have little role i n the regulation of prey numbers at natural densities i n this community. One potential response of predators I was not able to measure was predation rate. Individual rates of consumption may have increased i n response to increased prey density. Such a change may have been manifested i n increased predator growth rates, but my data were not adequate for such a test. Grazers were represented by only a few taxa i n this oligotrophic montane stream. There was no evidence for differences i n densities of grazers related to treat-ment. The only likely mechanisms that might have produced a response among graz-ers were i) altered microhabitat, ii) fertilization by nutrients leached from C P O M or iii) reduced predation pressure. The effects on grazers of species other than predators would be impossible to predict (Yodzis 1988) given our present state of knowledge. Indirect effects among community members may be subtle and take several generations to be measurable (Bender et al. 1984, Davidson et al. 1984, Carpenter 1988). The magnitude of a response should attenuate with increasing number of 109 trophic interactions from the target of the manipulation (McQueen et al. 1986). Most of the effects of C P O M additions on non-detritivore taxa were small, but in spite of the l imited power of the experiment, i.e., only two replicates per treat-ment (Toft and Shea 1983, Rotenberry and Wiens 1985), some indirect effects were apparent as a consequence of whole-leaf detritus additions. Increased densities of collectors with increased C P O M input supports the hy-pothesis that populations of F P O M feeders are facilitated by the processing of de-tritus by CPOM-detr i t ivores . That the response was detectable within two months, on the spatial scale of the experimental stream channels, demonstrates the intensity and importance of this interaction. The evidence for seasonal food hmitation of de-tritivores implies as a consequence, i n this "bottom-up" interaction, that collectors are also likely to be seasonally food limited i n this montane stream. 110 Chapter Six FOOD, HABITAT, OR BOTH?: MACROINVERTEBRATE USE OF LEAF ACCUMULATIONS IN STREAMS. Many organisms live on or i n their food, so that a single substrate serves as two resources. In such cases the lack of functional independence makes it diffi-cult to determine the importance of each resource type. For example, chrysomelid beetle larvae which consume l i ly pads not only reduce food, but deplete suitable habitat, the latter having a greater effect on population density (Juliano 1988). Thus the different functions of a single resource may affect different components of demography. Natural leaf packs i n streams serve as food and habitat to macroinvertebrates, but the relative importance of each function is difficult to isolate. Leaf packs provide food to many species, either directly or by accumulating fine detrital particles (Short and Masl in , Chapters 1, 4, 5). It has been suggested that leaf packs also serve as a component of habitat complexity which some species of invertebrates may occupy irrespective of food value (Egglishaw 1964). Thus, leaf packs provide shelter from direct current, space for settlement, and perhaps refuge from predators. A second issue is the influence of the larger scale habitat, e.g., riffles versus pools, on the use of leaf packs. Many studies of distribution of macroinvertebrates i n streams have sought to isolate the role of physical characteristics such as current velocity and mineral substratum from detrital abundance (Minshall and Minshall 1977, Rabeni and Minshall 1977, Reice 1980, Culp et al. 1983) but have all been done i n riffles. Since a large proportion of detrital retention is in pool or backwater habitats (Cummins et al. 1980, Speaker et al. 1984, Petersen et al. 1989), use of coarse particulate detritus by detritivores i n this habitat should not be ignored. Some species of C P O M consumers are actually more abundant i n pool habitats 111 rather than i n riffles (e.g., Lepidostoma quercina - Grafius and Anderson 1979). In addition, H u r y n and Wallace (1987) showed that while pools make up a smaller proportion of stream bottom, the areal production was actually higher i n pools than riffles i n an Appalachian mountain stream. Laboratory experiments have shown that macroinvertebrates prefer "condi-tioned" leaves, i.e, leaves from which soluble polyphenolic compounds have been leached and on which microbial growth has begun. To examine the influence of leaf conditioning on colonization by macroinvertebrates i n the field, I used leaf packs which had been incubated i n laboratory aquaria to compare with previously uncon-ditioned leaf packs and with artificial leaf packs made of polyester cloth. In this study I set out to test the following hypotheses. First , if habitat use (rif-fles versus pools) is independent of food sources, there should be habitat differences i n the processing rates of detritus and i n the density of consumers. Secondly, many organisms may use leaf packs as refuge rather than (or i n addition to) food; if so then I should be able to discriminate the use of leaf packs by supplying artificial leaf packs having no food value. Finally, to test the influence of soluble polyphenolics i n leaves on colonization rate I compared leaf packs cultured i n the laboratory against previously unconditioned leaf packs. M E T H O D S A N D M A T E R I A L S Study site These experiments were done i n Mayfly Creek, a second order stream i n the University of Bri t ish Columbia Research Forest (described i n Chapters 2 and 4). The riffles and pools used as sites for this experiment were immediately upstream of the confluence of Mayfly and Jacob's Creeks. Average daytime temperature of Mayfly Creek was approximately 11 C with a range of 6 to 16.5 C between 27 M a y and 16 July 1987. 112 Leaf pack manipulations Construction of leaf packs followed the description of Petersen and Cummins (1974). Alder leaves (Alnus rubra) were collected from the ground shortly after abscission i n the autumn of 1986, air-dried i n the laboratory, sealed i n plastic bags, and stored i n a freezer (-20 C) unti l use. Groups of leaves (10 - 15 leaves) were weighed to 5.0 g ( ± 0.025 g), fastened together with 3 small pieces of monofilament nylon through the leaves, and loosely tied to the upstream face of concrete bricks. Leaf packs were then randomly assigned to treatments. Leaf packs were placed i n the stream on five dates beginning 27 M a y 1987, and they were all retrieved on 16 July. This design differs from the usual practice of starting all the leaf packs on the same day and collecting them on different dates. The design used here means that there is no confounding of the data on animal numbers found i n the leaf packs due to life cycle phenology, changes i n population size between dates, and the recent history of the environment (Ciborowski and Chfford 1984). Art i f ic ial leaves were made from light brown polyester cloth cut into shapes similar to those of alder leaves. The edges of these cloth leaves were sealed by passing the edges through a flame to prevent fraying. These "leaves" were gathered i n groups of 12 and fastened together the same way as the real leaf packs, and tied with monofilament to the upstream face of concrete bricks. The grouping of 12 leaves was chosen to approximate the surface area provided by 5.0 g of alder leaves. Leaf packs were placed i n two riffles and two pools of Mayfly Creek. O n each of five dates, three (replicate) leaf packs were added to each of the four stream sections. O n three dates (not including 27 M a y or 14 July) three artificial leaf packs were added to each of the two pools and six packs added to each of the riffles. On 14 July only three artificial leaf packs were added to each of the riffles and pools. The date of placement was identified by a piece of coloured surveying tape attached to the brick. 113 Leaf packs for the test of conditioning on colonization were prepared i n the same manner as the other leaf packs. Six alder leaf packs were placed i n aquaria i n an environmental chamber (15 C) on each of 16 June, 30 June and 7 July. The water for the aquaria was collected i n the field, filtered through a 63 / i m sieve, and diluted 1:1 with dechlorinated water. The water was aerated continuously and the water was replaced twice a week. Leaves collected from Mayfly Creek were rinsed and added to each aquarium to provide a further microbial inoculum for the leaf packs. The leaf packs i n the aquaria were allowed to begin microbial conditioning i n the aquaria, and on 14 July three leaf packs for each of the three initiation dates were placed i n each of the two riffles among the other leaf packs already i n place. These leaf packs are referred to as "preleached". O n 16 July 1987, all leaf packs were retrieved from the stream by lifting bricks and attached leaves with a 250 /xm mesh net positioned immediately downstream to catch any macroinvertebrates which left the leaf pack when it was disturbed. In the laboratory the associated invertebrates were washed off the leaves. The leaves were dried at 60 C for 24 h , then weighed on a Mettler balance to the nearest mg. Leaves were reduced to ash i n a muffle furnace and reweighed to determine ash-free dry mass. Animals associated with the leaf packs were washed through a 1.0 m m sieve and those retained on the sieve were identified and counted. Counts of animals passing through the 1.0 m m sieve in 14 samples showed that no specimens of the larger detritivore taxa passed through. Since the animals retained on the 1.0 m m sieve are the major consumers of coarse detritus during the summer, the use of a large mesh sieve did not bias the results. The data for numbers of macroinvertebrates per leaf pack were transformed by log1Q(n + l). To test for statistical assumptions of A N O V A , the data were first fitted to the complete General Linear Model [for unbalanced A N O V A designs, SAS (1985)] 114 and the residuals were compared for deviation from a normal distribution using the Shapiro-Wilk statistic. Those sets of data meeting the appropriate assumptions for A N O V A are presented here. The effect of riffle or pool location is called a habitat effect, while real and artificial leaf pack comparisons are termed a food effect. The numbers of colonists on leaf packs within 48 h were very small and variable, so comparisons among preleached leaf packs, unleached leaf packs, and artificial leaf packs were made with Kruskal-Wallis tests (SAS 1985). The numbers of macroinvertebrates on leaf packs of different degrees of con-ditioning and i n the different habitats were compared based on number per leaf pack, and number per gram of leaf pack remaining. There are several assumptions underlying the use of either measure of animal density (see Discussion), but both measures are commonly used i n the literature. R E S U L T S Leafpack decomposition rates The data for alder leafpack decomposition were fitted to least square linear regressions (Table XII ) . A negative exponential model fitted to the same data ac-counted for only 65.6% of the total variance. The data for dry weight and ash-free dry weight gave similar results so the analysis for dry weight is shown i n Table XI I . The A N C O V A (Table XII) showed there were no significant differences i n decompo-sition rates between riffles and pools, or between upstream or downstream sections of the stream (Figure 27), as indicated by lack of significant interaction with the covariate (number of days i n the stream). The least square means for the mass of leaf packs from riffles were slightly less than those for pool leaf packs but not sig-nificantly so. Since the true units of replication for these comparisons are the pairs 115 T a b l e X I I — Summary A N C O V A table for decomposition rates of alder leaf packs, comparing the rates i n riffle and pool areas. Total R 2 = 0.926. Source d.f. S.S. F P Habitat 1 0.079 0.60 0.44 Upstream/Downstream 1 0.083 0.63 0.43 Days 1 81.241 617.49 < 0.0001 Habitat X Days 1 0.256 1.95 0.17 U p / D o w n X Days 1 0.042 0.32 0.57 Error 51 90.697 of riffles or pools, the test using all leaf packs is biased toward finding a difference, so the possibility of a type II error is very low. The overall regression (irrespective of pool or riffle, but not including preleached leaf packs) of leaf pack mass (in g) over time was mass = 4.311 — 0.071 days. This means that within the first two days the loss due to handling and leaching was 0.831 g, or 16.6% of the initial dry mass. The R 2 for the regression was 0.90. In order to compare decomposition rate with other published values I calculated the exponent of the decay curve (in spite of its inferior fit to the data) to be -0.0469. Macroinvertebrates For all taxa tested, with the exception of Baetis (p > 0.18), there were sig-nificant (p < 0.01) treatment effects i n the comparisons of natural to artificial leaf packs (Table XIII) . There were more individuals found on the natural leaves on most dates. There were also significant differences between leaf packs placed on different dates ( P < 0.05). In most cases the analysis of numbers per gram of leaf remaining (excluding artificial leaf packs) gave similar results as the analysis of number per leaf pack; exceptions wi l l be noted below. Malenka nymphs (including M. californica and M. cornuta) were found almost exclusively on natural leaf packs i n the riffles (Figure 28). The other 3 combinations 116 Days in S t r eam F i g u r e 27 — Regressions of alder leaf pack mass remaining over time in Mayfly Creek. The number of leaf packs recovered was 57, which are shown here; some points are hidden. 117 T a b l e X I I I — Summary of A N O V A [General Linear Models - SAS (1985)] for number of individuals per leaf pack. The tests were for the effect of real (red alder) versus artificial leaf packs i n replicate riffles and pools. Effect Malenka spp. Food Habitat Upstream/Downstream Days Food * Days Habitat * Days Lepidostoma roafi Food Habitat Upstream/Downstream Days Food * Days Habitat * Days Total Detritivores Food Habitat Upstream/Downstream Days Food * Days Habitat * Days Baetis spp. Food Habitat Upstream/Downstream Days Food * Days Habitat * Days non - detritivores Food Habitat Upstream/Downstream Days Food * Days Habitat * Days df F P 1 27.3 < 0.0001 1 19.1 < 0.0001 1 0.8 0.38 4 6.3 < 0.0001 3 3.9 < 0.02 4 2.9 < 0.03 1 82.7 < 0.0001 1 6.6 < 0.02 1 7.0 < 0.01 4 9.2 < 0.0001 3 13.0 < 0.0001 4 1.6 0.18 1 78.4 < 0.0001 1 5.6 < 0.03 1 3.9 0.05 4 9.9 < 0.0001 3 8.4 < 0.0001 4 0.2 0.93 1 1.7 0.19 1 5.3 < 0.03 1 0.4 0.55 4 2.5 < 0.05 3 5.6 < 0.002 4 0.7 0.61 1 37.3 < 0.0001 1 26.0 < 0.0001 1 1.2 0.28 4 8.2 < 0.0001 3 4.9 < 0.003 4 1.4 0.24 118 were similar by the absence of more than a few nymphs. The difference i n the pattern of colonization of the 4 treatments accounts for the significant effect of habitat (Table XIII) and interactions of food X day and habitat x day. For the number of Malenka nymphs per gram dry mass of leaf pack remaining there is a monotonic increase as the colonization period increased for the natural leaf packs from riffles (Figure 28). Larvae of the caddisfly Lepidostoma roafi were essentially absent from the ar-tificial leaf packs (Figure 29, Table XIII) . For four of the five colonization periods there were more larvae associated with natural leaf packs i n pool areas compared with riffles, however, the habitat effect was not significant when numbers per gram of leaf were considered (excluding artificial leaves, JFI,43 = 1.86, P > 0.15). The difference i n the shape of the colonization pattern between treatments resulted in a significant interaction of food x day. For this species there were also differences between the pairs of riffles and pools. As with Malenka there was a consistent increase i n the numbers per gram of leaf remaining over the experimental period (Figure 29). The total numbers of large detritivores are compared i n Table XII I (see also Figure 30). Large detritivores included L. roafi, Malenka spp., as well as second year Despaxia augusta, the caddisflies L. unicolor, L. cascadense, Ecclisomyia conspersa, Psychoglypha spp. and Ecclisocosmoecus scylla, and a few rarer taxa. There were significant effects of food, habitat, days, and food x days (Table XIII) . There were significant differences between the pairs of riffles and pools. As for Malenka and L. roafi there were few detritivores associated with artificial leaf packs. For all dates the combined numbers of large detritivores were higher on riffle leaf packs. The mayfly Paraleptophlebia debilis showed a significant effect of habitat, while its congener P. temporalis had no significant differences in numbers i n riffles or pools (Table XIII , Figure 31). For both taxa, numbers per leaf pack were higher on packs 119 F i g u r e 28 — Malenka nymphs per leaf pack, alder and artificial, over period of residence in situ. 120 o ro M— r o <U <u Q. <u _Q E 3 6-1 4 2 0 O — r e a l , riffle • — r e a l , poo l A — a r t i f i c i a l , riffle A — a r t i f i c i a l , poo l Lepidostoma roafi r o E r o i _ i _ <u Q. E 3 4 34 2 1 0 0 10 20 30 40 Days in S t r eam 50 F i g u r e 29 — Lepidostoma roafi nymphs per leaf pack, alder and artificial, over period of residence in situ. 121 To ta l Detr i t ivores i 1 1 1 T r 0 10 20 30 40 50 Days in S t r eam F i g u r e 30 — Numbers of large (> 1.0 mm) detritivores per leaf pack, over period of residence in situ. 122 placed i n pools. Both species also had significant differences i n the numbers i n the replicate pools and riffles, although this effect was not significant when numbers per gram of leaf were tested. The numbers on artificial leaf packs were very low, but showed an increase with increasing colonization period. Baetis larvae were more abundant i n riffles (Figure 32). Although the food effect was not significant by itself, there was a food x day interaction (Table XIII) . The numbers of these mayflies on artificial leaf packs increased more slowly than on real leaves. By 16 days of colonization densities on artificial leaves were greater than those on real leaf packs i n their respective habitat. The combined totals of species other than C P O M consumers collected from leaf packs (including Paraleptophlebia and Baetis) were significantly different for all main effects, except for differences between habitat replicates and for the food X date interaction (Table XIII) . The food effect alone explained 37% of the total variance i n the data. Colonization Colonization was taken as the net rate of change i n the number of individuals per leaf pack over the period of the experiment. The numbers of large macroinver-tebrates colonizing (during the first 2 days) artificial and real leaf packs i n pools and riffles, and real leaf packs which had been preleached for different periods of time were compared for a number of taxa (Figure 33). In most cases there were few colonists on artificial leaf packs. There were significant differences i n colonization for P. debilis, Baetis, and the total number of large non-shredders (Kruskal-Wallis, 6 d.f., P < 0.01). In all cases the differences were between real leaf packs i n riffles and the others, either artificial or i n the pool habitats. Colonization rates of detritivores on leaves within a treatment were similar during the first 30 days (Figures 28 - 32). Colonization declined or became negative 123 u ro Q_ CL <u _Q E O — r e a l , riffle 7 _ • — r e a l , poo l 64 Paraleptophlebia debilis 4 2 A — a r t i f i c i a l , riffle #^ A — a r t i f i c i a l , poo l / \ / \ / \ \ \ . / / / / / \ \ \ \ \ \ \ / \ \ -O —T~ Paraleptophlebia 0 10 20~~ 30 40 50 Days in S t r eam F i g u r e 31 — Numbers of Paraleptophlebia debilis and P. temporalis per leaf pack. 124 Days in S t r eam F i g u r e 32 — Numbers of Baetis spp. per leaf pack. 125 Total Large Detritivores 0 Total Large non-Detritivores <u " D - D rjz o 0 0 ^ r? r? OJ CsJ i—I h- LJ_ EE _ r ° ^ H " O " D r o J-* re Cr_ «u <v QJ •- Q_ •-_ -c _c _c .!=! _ — u u u — '-tr f o r o r o r o '+J ?'•+-> ^ <u o> <v u i -_ J _ i _ i < c t r < F i g u r e 33 — Numbers colonizing leaf packs within two days of placement into stream. The number of colonists on real and artificial leaf packs, as well as the number on preleached leaf packs, are given. 126 after 30 days. I have no estimates of turnover rates of individuals, but this is not relevant to this paper. D I S C U S S I O N That leaf packs of alder were used to a similar degree i n both riffle and pool habitats is supported by two lines of evidence. First , the differences i n decomposi-tion rates of leaf packs were indistinguishable between habitats. Second, the total densities of detritivores were similar on leaf packs i n either habitat, with the ex-ception of the higher value i n riffles for leaf packs which had been i n the stream for 30 d. The similarities of leaf pack use i n both habitats suggests that patches of resources are readily discovered and consumed i n either habitat. Since the storage location of detritus varies temporally as a result of changes i n discharge (Cummins et al. 1980, Speaker et al. 1984) and given that C P O M was i n short supply i n Mayfly Creek (Chapter 4), an organism could thus benefit from being flexible i n its choice of habitat. Leaf decomposition studies are usually done with leaf packs i n riffles only (e.g., M c A r t h u r et al. 1988) and usually during autumn and winter. M y results suggest that the difference i n decomposition rates between pools and riffles are minor. Rates of leaf decomposition are also usually positively related to nutrient concentrations i n the water (Howarth and Fisher 1976, Elwood et al. 1981). The oligotrophic nature of Mayfly Creek, low densities of detritivores, and lower temperatures, might be expected to result i n relatively slower rates of decomposition. O n the other hand the shortage of food for detritivores i n this stream could mean that faster decay rates is a relative index of food limitation. Although natural leaf pack accumulations are scarce i n Mayfly Creek during summer, they are present because the green leaves are not flushed out of riffles, as they would be by autumn and winter freshets (Chapter 2). 127 The low numbers of colonists on artificial leaf packs suggest that microhabitat is not an important role of C P O M accumulations and that, at least for some species, accumulations of leaves are not used for refuge. During summer, single leaves (i.e., offering no refuge) were often found submerged and with relatively large numbers of detritivores visible on the surface. Lepidostoma roafi and Ecclisomyia conspersa (Trichoptera) larvae were both very conspicuous on the surface of these isolated leaves. In many cases there was a gradual increase i n numbers of insects on the artificial leaf packs. This might indicate that the artificial packs slowly accumulate F P O M , periphyton, or small amounts of C P O M . A response such as that shown by Baetis, for which leaf packs are not directly food, suggests that leaf packs probably serve another function such as a site for the collection of F P O M and periphyton or refuge. The relative absence of most taxa from artificial leaf packs shows that food and habitat are not independent. Accumulations of C P O M i n streams serve primarily as food; alternative functions, such as refuge or substrate, are either unimportant or masked by the value as food. A n experimental manipulation of food (straw) and substratum i n a small pond by Street and Titmus (1982) showed that substratum accounted for 14 x the variance associated with macroinvertebrate densities explained by the food addition. The only other study to examine the effect of C P O M as microhabitat only, showed that macroinvertebrates did not colonize substrate trays with fragments of rubber, while addition of natural detritus resulted i n high rates of colonization (Egglishaw 1964). A n alternative explanation for the low numbers of colonists i n my experiment is that the artificial leaves used in this experiment were not reasonable mimicks of real leaves (perhaps even toxic). The increasing numbers over time, particularly for non-detritivore species, suggests that this is not the case. There were some taxonomic differences i n the use of leaf packs i n the two habitats. Larvae of Malenka were mainly restricted to riffles, whereas Lepidostoma 128 roafi occurred i n about equal numbers on leaf packs i n either habitat. Nymphs of the mayfly Paraleptophlebia were much more abundant on pool leaf packs than i n riffles. The factors leading to differences i n species use of habitat are unknown for these taxa; however, the primary result is that densities of all insects combined did not differ significantly between habitats. In addition there was no effect of the habitat type on use of artificial leaf packs. I have shown earlier (Chapter 4) that addition of supplemental whole leaves to stream channels resulted i n increased densities and growth rates of detritivores. One hypothesis which might explain the results of the detritus-supplement experiment was an increase i n suitable microhabitat and refuge from predation. The results from this leaf pack experiment suggest that the microhabitat hypothesis is unlikely to account for much of the density increase i n the food supplement experiment (Chapter 4). Several studies have found the highest densities and highest species richness i n leaf accumulations, relative to other stream substrata (e.g., Mackay and Kalff 1969), and this may be primarily a function of food supply. The presence of soluble phenolic compounds i n real leaves may reduce the at-tractiveness of leaf packs to colonists. Allowing leaf packs to leach soluble organics i n the laboratory prior to positioning i n the field had no significant effect on any taxon considered. There were higher numbers of colonist detritivores on preleached leaf packs, but the effect of leaching (and some microbial growth) was subtle and probably complex. Perhaps the presence of soluble polyphenolics does not influ-ence colonization. Alternatively, if leaching occurs quickly enough, then a 2-day colonization period may be too long to see an effect of preleaching. Data on leaf pack colonization by macroinvertebrates are usually expressed as numbers per leaf pack or numbers per gram of leaf remaining. W h i c h of these measures to use depends on some assumptions regarding behaviour and changes i n the quality and quantity of material in a leaf pack which have never been explicitly 129 considered. As decomposition proceeds the remaining leaf material wi l l become increasingly composed of lignified, refractory components such as vascular bundles (Barlocher 1983), so the overall quality (including microbial growth) presumably in -creases to overcome the overall reduction i n leaf pack mass. The increasing numbers of large shredders per gram of leaf pack remaining can come about i n at least two ways. First , because numbers on a leaf pack represent the net result of immigration and emigration processes, continued increase per gram could represent a resource i n which increasing quality offsets the diminishing quantity. Second, there could be a lag time i n the response of individuals leaving the leaf pack due to the decline i n the quality or quantity of food. The distinction is relevant to how organisms choose patches of food relative to food quality and the numbers of potential competitors already there. The "rules" an individual uses to choose to settle i n a patch may be different from those used to "decide" to emigrate. This asymmetry may lead to apparently increased numbers per mass of food remaining i n spite of a decline i n quality or quantity. A comparison of the number per leaf pack and the number per gram of leaf remaining may allow some disentangling of these two processes. It appears that leaf packs i n streams attract animals primarily for the food value. The lack of colonization of artificial leaf packs plus observations of animals on single leaves, suggest that stream macroinvertebrates do not seek habitat inde-pendently of their food supply. While natural leaf packs provide food, they may incidentally function as refuge. To test the latter hypothesis would require manipu-lations of predators, detritivores, and the abundance of C P O M to examine mortality rates. Nevertheless there is no evidence that provision of substratum alone accounts for the colonization of leaf accumulations. 130 Chapter 7 GENERAL DISCUSSION In this thesis I have shown that detritivores of a small montane stream are food l imited and that food supplementation results i n increased densities and growth rates of common species (Chapter 4). The cause of the density response to added whole-leaf detritus appears to be a true food effect and probably not an increase i n habitat (Chapter 6). Additions of coarse detritus resulted i n increased densities of fine particle consumers (Chapter 5), providing what seems to be the first exper-imental field evidence of the important assumption of this trophic connection i n stream ecosystem theory. During the past 25 years, evidence has accumulated that lotic animals i n trophic levels other than detritivores may also experience food hmitation. Consumers of periphyton are capable of resource depression at natural densities (McAuliffe 1984, Power et al. 1985, Hart 1987, H i l l and Knight 1987, Lamberti et al 1987) and reach higher densities and/or faster growth rates when algae are fertilized. Individual grazers have been shown to grow faster when densities are experimentally reduced (Hart 1987, H i l l and Knight 1987). Collectors of fine particulate detritus also ap-pear sometimes to be food limited (see Chapter 5). Salmonids are evidently food limited based on positive correlations between stream productivity and fish size and/or density (Murphy et al. 1981) and on the results of various food enhance-ment experiments (Warren et al. 1964, Mason 1976). The evidence available from the studies cited above suggests that productivity at each trophic level i n many streams may be limited by food supply. One difficulty with showing population limitation by resource supplementation is the possibility of life history compensation (Sibly and Calow 1987). Most studies consider only a limited part of the life cycle of the organism of interest so that 131 short-term positive responses to food addition (e.g., increased density and increased growth) may not necessarily indicate food hmitation of population size. In my food addition study the summer-emerging species had already gone through the winter period early i n their life cycle during which mortality rates are high. Even though densities may have been reduced during autumn and winter, these summer-emerging species stil l were food limited. The possibility remains that higher densities may be compensated for by changes i n rates of mortality, for instance, if refugia from winter floods are l imited. This problem remains outside the scope of my study and I suspect it wi l l be difficult to address i n natural systems. Implications for Community Structure Perspectives of stream community structure have shifted i n the past two decades from a predominantly abiotic, environmentally dominated view (Hynes 1970) to one which includes a rich diversity of processes (e.g., papers i n Barnes and Minshal l 1983, Fontaine and Bartell 1983). Current issues include assessment of the role of resource l imitation, vertebrate and invertebrate predators, disturbance, periphyton-consumer interactions and geomorpholgy. To date, most evidence for the role of vertebrate predators suggests little effect on densities of benthic lotic prey (AUan 1982, Flecker 1984, Reice and Edwards 1986), although there are some studies showing some definite effects (Hemphill and Cooper 1984, Power et al. 1985). The complex nature of stream bottoms means that many stream fishes (notably salmonids) are relegated to feeding on drifting in-vertebrates which may be individuals i n excess of carrying capacity (Waters 1972) and suggests that mortality to fish predation may be compensatory. This "bottom-up" view of stream communities implies that fish predators have little influence i n determining population densities of prey. The role of invertebrate predators is essentially unknown. Although productivity within a guild or trophic level may be 132 food l imited, this does not imply that predators are unimportant i n affecting com-munity structure. Size selectivity, frequency dependence and density dependence of predation, as well as prey behaviour and morphology (Molles and Pietruszka 1983) wi l l all likely be influential factors i n the arrangement of stream communities. The role of disturbance is still relatively difficult to estimate and to factor into concepts of lotic community organization (Fisher 1987, Hildrew and Townsend 1987). In streams disturbance is generally equated with the physical effects of floods. There is little evidence of a true rarifying effect of floods on communities, but there is now one study which has shown overwinter depression of numbers proportional to flood frequency for a northern Californian stream (McElravy et al. 1989). Flooding may have more of an influence on benthic populations by scouring and flushing of algal and detrital resources, rather than by direct mortality of relatively mobile invertebrates (Anderson and Lehmkuhl 1968, Hildrew and Townsend 1987). Disturbance, predation, parasitism and other processes are likely to have im-portant roles i n community organization. Nevertheless the influence of the above mechanisms does not negate the function of food supply affecting population densi-ties and productivity according to the "bottom-up" view, contra Bowlby and Roff (1986). Evolution of life cycle phenology In many papers dealing with aquatic detritivores, the timing of life cycles is claimed to be "synchronized" with the seasonal pattern of leaf litter inputs (e.g, Peterson and Cummins 1974, Hanson et al. 1984, Stout and Taft 1985). There are contradictions i n terms of what the exact timing should be, i.e., maximizing the growth and survival of early instar larvae (Petersen and Cummins 1974) or maximizing growth of nearly mature larvae (Hanson et al. 1984). In such a claim is the implicit recognition that seasonality of this resource may affect the selection of life history traits and timing. The interesting aspect of this is that the periods 133 when food supply is l imiting are when density dependent feedback can affect the evolution of life history traits (Hastings 1984). Even if seasonal variation i n resource abundance has influenced the evolution of life cycle timing, it is likely to have been a response to life history "bottlenecks" imposed by limited food supply, rather than by peak abundance of food. In most invertebrates the very young stages are most susceptible to food short-age, since they lack substantial reserves that might sustain them through short periods of resource depression (e.g., Neil l and Peacock 1980). Thus, even though the timing of life cycle events may favour maximizing growth rates of nearly mature larvae (for instance, the Zapada spp. i n Mayfly Creek), the early stages may suffer higher rates of mortality than if they developed later. O n the other hand, some species, for example Potamophylax cingulatus (Otto 1974) may pass through their early instars during the "flush" of detrital resources and then experience food l i m -itation of growth rates when they elaborate l ipid reserves during their final instar. The potential for intraspecific competition means that individuals which are larger or more efficient wi l l have a selective advantage when resources are scarce; when re-sources are abundant there wi l l be no density dependent selection based on resource use. M a n y other factors can affect the evolution of life history phenology, including timing of disturbance events, presence of competitors and historical constraints. To ascribe any single factor as responsible for particular life cycle patterns would be pure speculation, and it is unlikely that single factors ever operate i n isolation. More often there are tradeoffs imposed by the interaction of several effects. Future Research Directions There are several obvious directions for research following from the work I have presented. The first two, which are intimately related, are the mechanisms of competition and means of coexistence; both were alluded to in chapter four. Several aspects of competition require examination. First , the actual behavioural 134 interactions (or perhaps lack thereof) need to be explored. There are likely to be species- and size-biased interactions among individuals. Evidence of a size bias between emigrant and resident Zapada cinctipes (Chapter 4) suggests size dependent interactions. Second, the demographic consequences of competition on each taxon need to be quantified. Finally, the nature of the density dependence of competition should be quantified and related questions of rate of feedback, seasonal changes i n intensity, etc. should be addressed. The issue of coexistence remains unresolved. As stated, microhabitat or food specialization are unlikely to be large contributors to coexistence i n this system. While some species are separated temporally to some degree, periods of resource use overlap extensively, and pairs of congeners have similar hfe cycle phenology. The relative importance of resource limitation, predation and disturbance to the determination of stream macroinvertebrate densities is an elusive problem. The study of disturbance effects is complicated by the large scale of the process, i.e., whole watersheds, and wil l be difficult, if not impossible, to adequately manipulate experimentally. The test of the effects of floods may have to wait unti l long term detailed hfe table data are available, since manipulation of the factors associated with spates on an appropriate spatial and time scale is unlikely (e.g., Fisher 1987). The generality of the food limitation model could be tested i n other stream systems. For example, are detritivores i n deciduous forest streams ever food l im-ited? - some opinions suggest not (Cummins et al. 1989, S.R. Reice - personal communication). Freshwater limnetic organisms are more strongly food and nutri-ent l imited under oligotrophic conditions (McQueen et al. 1986, McQueen and Post 1988, Neil l 1988). Perhaps stream productivity hkewise leads to patterns of food web properties among stream communities similar to those of lakes. A n extension of the food limited detritivore model might be soil litter organ-isms. The soil system is analogous to the stream community i n many ways, yet 135 i n my search of the literature on soil litter organisms, I have been unable to find the suggestion that soil detritivores may be food limited at any part of the year. The literature on terrestrial detritus has apparently focussed on rates and channels of cycling and mineralization processes of detrital materials i n the same way the stream literature has been focussed (Dickinson and Pugh 1974, Swift et al. 1979, Hunt et al. 1987). Thir ty years ago, Hairston et al. (1960) proposed that detritivores should usu-ally be food Hrnited based on the observation that detritus does not accumulate. Their premise may be easily challenged for most ecosystems, however streams do appear to have a long-term average quantity of detritus (Cummins et al. 1983). If Hairston et aL's hypothesis was refined to indicate some measure of quality then it may be more general. This detritus hypothesis of Hairston et al. has been mostly ignored i n favour of their assertion about predator l imitation of herbivore popu-lations (Oksanen et al. 1981, Oksanen 1988, McQueen et al. 1986, Fretwell 1987). Nevertheless, as discussed i n earlier chapters there is evidence that detritivores i n many systems are sometimes food limited. Indeed, Pearson and Rosenberg (1978, 1987) argued that gradients of organic enrichment may underlie all marine ben-thic fauna! densities and distributions. While detritivores and collectors were food l imited i n Mayfly Creek, the conditions leading to food limitation i n streams are unknown, e.g., oligotrophy, seasonal fluctuations of food abundance, forest type, etc. Determination of the generality of detritivore food limitation i n streams and other ecosystems wil l have to await further tests. 136 LITERATURE CITED A l l a n , J . D . 1982. The effects of reduction i n trout density on the invertebrate community of a mountain stream. Ecology 63:1444-1455. A l l a n , J . D . 1984. Hypothesis testing i n ecological studies of aquatic insects, pp. 484-507 In : Resh, V . H . and D . M . Rosenberg (Eds.). The Ecology of Aquatic Insects. Praeger, Toronto. Alongi , D . M . and K . R . Tenore. 1985. Effect of detritus supply on trophic relation-ships within benthic food webs. I. meiofauna-polychaete (Capitella capitata (Type I) Fabricius) interactions. Journal of Experimental Marine Biology and Ecology 88:153-166. Anderson, N . H . and D . M . Lehmkuli l . 1968. Catastrophic drift of insects i n a woodland stream. Ecology 49:198-206. Anderson, N . H . and J .R. Sedell. 1979. Detritus processing by macroinvertebrates i n stream ecosystems. Annual Review of Entomology 24:351- 377. Anderson, N . H . , J .R . SedeU, L . M . Roberts and F . J . Triska. 1978. The role of aquatic invertebrates i n processing of wood debris i n coniferous forest streams. American Midland Naturalist 100:64-82. Anderson, N . H . and J . L . Wold. 1972. Emergence trap collections of Trichoptera from an Oregon stream. Canadian Entomologist 104:189-201. Andrewartha, H . G . and L . C . Birch . 1954. The Distribution and Abundance of Animals . University of Chicago Press, Chicago. Arsuffi , T . L . and K . Suberkropp. 1988. Effects of fungal mycelia and enzymatically degraded leaves on feeding and performance of caddisfly (Trichoptera) larvae. Journal of the North American Benthological Society 7:205-211. Barlocher, F . 1983. Seasonal variation of standing crop and digestibility of C P O M i n a Swiss Jura stream. Ecology 64:1266-1272. Barlocher, F . and B . Kendrick. 1973. Fungi and food preference of Gammarus pseudolimnaeus. Archiv fiir Hydrobiologie 72:501-516. Barlocher, F . and B . Kendrick. 1975. Assimilation efficiency of Gammarus pseu-dolimnaeus (Amphipoda) feeding on fungal mycelium or autumn-shed leaves. Oikos 26:55-59. Barnes, J .R . and G . W . Minshall (Eds.) 1983. Stream Ecology: application and testing of general ecological theory. Plenum Press, N Y . 137 Baumann, R . W . , A . R . Gaufin and R . F . Surdick. 1977. The Stoneflies (Plecoptera) of the Rocky Mountains. Memoirs of the American Entomological Society 31:1-208. Bender, E . A . , T . J . Case and M . E . Gi lp in . 1984. Perturbation experiments in community ecology: theory and practice. Ecology 65:1-13. Benfield, E . F . and J .R. Webster. 1985. Shredder abundance and leaf breakdown i n an Appalachian Mountain stream. Freshwater Biology 15:113-120. B i r d , G . A . and N . K . Kaushik. 1981. Coarse particulate organic matter i n streams, pp. 41-68 In: Lock, M . A . and D . D . Will iams (Eds.). Perspectives i n Running Water Ecology. Plenum, N Y . Bowlby, J . N . and J . C . Roff. 1986. Trophic structure i n southern Ontario streams. Ecology 67:1670-1679. Boyce, M . S . and D . J . Daly. 1980. Population tracking of fluctuating environments and natural selection for tracking ability. American Naturalist 115:480-491. Brown, J.S. 1989. Coexistence on a seasonal resource. American Naturalist 133:168-182. Carpenter, S.R. (Editor). 1988. Complex Interactions i n Lake Communities. Springer-Verlag, N Y . Carpenter, S.R., J . F . Kitchel l and J .R. Hodgson. 1985. Cascading trophic interac-tions and lake productivity. BioScience 35:634-639. Cather, M . R . and A . R . Gaufin. 1976. Comparative ecology of three Zapada species of M i l l Creek, Wasatch Mountains, Utah (Plecoptera: Nemouridae). Ameri-can Midland Naturalist 95:464-471. Chesson, P . L . 1986. Environmental variation and the coexistence of species, pp.240-256 In: Diamond, J and T . J . Case [Eds.], Community Ecology. Harper and Row, N Y . Ciborowski, J . J . H . and H . F . Clifford. 1984. Short-term colonization patterns of lotic macroinvertebrates. Canadian Journal of Fisheries and Aquatic Sciences 41:1626-1633. Clifford, H . F . 1969. Limnological features of a northern brown-water stream, with special reference to the life histories of the aquatic insects. American Midland Naturalist 82:578-597. Colwell , R . K . 1974. Predictability, constancy, and contingency of periodic phenom-ena. Ecology 55:1148-1153. 138 Corinell, J . H . 1983. O n the prevalence and relative importance of interspecific competition: evidence from field experiments. American Naturalist 122:661-696. Connors, M . E . and R . J . Naiman. 1984. Particulate allochthonous inputs: relation-ships with stream size i n an undisturbed watershed. Canadian Journal of Fisheries and Aquatic Sciences 41:1473-1484. Culp J . M . , S.J. Walde and R . W . Davies. 1983. Relative importance of substrate particle size and detritus to stream benthic macroinvertebrate microdistribu-tion. Canadian Journal of Fisheries and Aquatic Sciences 40:1568-1574. Cummins, K . W . 1973. Trophic relations of aquatic insects. Annual Review of Entomology 18:183-206. Cummins, K . W . 1974. Structure and function of stream ecosystems. BioScience 24:631-641. Cummins, K . W . and M . J . K l u g . 1979. Feeding ecology of stream invertebrates. Annual Review of Ecology and Systematics 10:147-172. Cummins, K . W . , R . C . Petersen, F . O . Howard, J . C . Wuycheck and V . I . Holt . 1973. The utilization of leaf Utter by stream detritivores. Ecology 54:336-345. Cummins, K . W . , J .R . Sedell, F . J . Swanson, G . W . Minshal l , S .G. Fisher, C . E . Cush-ing, R . C . Petersen and R . L . Vannote. 1983. Organic matter budgets for stream ecosystems: problems i n their evaluation, pp. 299-353. In: Barnes, J .R . and G . W . Minshall [Eds.]. Stream Ecology: application and testing of general ecological theory. Plenum Press, N Y . Cummins, K . W . , G . L . Spengler, G . M . Ward, R . M . Speaker, R . W . Ovink, D . C . M a -han, and R . L . Mattingly. 1980. Processing of confined and naturally en-trained leaf Utter i n a woodland stream ecosystem. Limnology and Oceanog-raphy 25:952-957. Cummins, K . W . , M . A . Wilzbach, D . M . Gates, J .B . Perry and W . B . TaUaferri. 1989. Shredders and riparian vegetation. BioScience 39:24-30. Cushing, J . M . 1986. Oscillatory population growth i n periodic environments. The-oretical Population Biology 30:289-308. Daniel , W . W . 1978. Applied Nonparametric Statistics. Houghton Miff l in , Boston. Davidson, D . W . , R.S. Inouye and J . H . Brown. 1984. Granivory i n a desert ecosys-tem: experimental evidence for indirect facilitation of ants by rodents. Ecol-ogy 65:1780-1786. DeAngeUs, D . L . 1975. Stability and connectance in food web models. Ecology 56:238-243. 139 Dickinson, C H . and G. J .F . Pugh (Eds.) 1974. Biology of Plant Litter Decomposi-t ion, Volumes 1 and 2. Academic Press, N Y . Drake, J . A . 1984. Species aggregation: the influence of detritus i n a benthic inver-tebrate community. Hydrobiologia 112:109-115. Eadie, J . M . and A . Keast. 1982. Do Goldeneye and Perch compete for food? Oecologia (Berlin) 55:225-230. Egglishaw, H . J . 1964. The distributional relationship between the bottom fauna and plant detritus i n streams. Journal of Animal Ecology 33:463-476. El l i s , R . J . 1975. Seasonal abundance and distribution of adult stoneflies of Sashin Creek, Baranof Island, southeastern Alaska. The Pan-Pacific Entomologist 51:23-30. Elwood, J . W . , J . D . Newbold, A . F . Trimble and R . W . Stark. 1981. The limiting role of phosphorus i n a woodland stream ecosystem: effects of P enrichment on leaf decomposition and primary producers. Ecology 62:146-156. Emlen, J . M . 1984. Natural selection and population density-feedback I. population hmitation by food shortage, predation, and self-damping factors. Theoretical Population Biology 25:41-61. Feller, M . C . 1977. Nutrient movement through western hemlock-western redcedar ecosystems i n southwestern Brit ish Columbia. Ecology 58:1269-1283. Feller, M . C . and J .P . Kimmins. 1979. Chemical characteristics of small streams near Haney i n southwestern Brit ish Columbia. Water Resources Research 15:247-258. Findlay, S .E .G . 1982. Effect of detrital nutritional quality on population dynamics of a marine nematode (Diplolaimella chitwoodi). Marine Biology 68:223-227. Fisher S .G. 1987. Succession, scale and hypothesis testing i n streams. Canadian Journal of Fisheries and Aquatic Sciences 44:689. Fisher, S .G. and G . E . Likens. 1973. Energy flow i n Bear Brook, New Hampshire: an integrative approach to stream ecosystem metabolism. Ecological Mono-graphs 43:421-439. Flecker, A . S . 1984. The effects of predation and detritus on the structure of a stream insect community: af ie ld test. Oecologia (Berlin) 64:300-305. Fontaine, T.D.,III and S . M . Bartell (Eds.)." 1984. Dynamics of Lotic Ecosystems. A n n Arbor Science, Michigan. Fretwell, S.D. 1972. Populations i n a Seasonal Environment. Princeton University Press, Princeton, N J . 140 Fretwell, S .D. 1987. Food chain dynamics: the central theory of population ecology? Oikos 50:291-301. Gee, J . H . R . 1988. Population dynamics and morphometries of Gammarus pulex L . : evidence of seasonal food limitation i n a freshwater detritivore. Freshwater Biology 19:333-343. Gee, J . M . , R . M . Warwick, M . Schaanning, J . A . Berge and W . G . Ambrose, Jr. 1985. Effects of organic enrichment on meiofaunal abundance and community struc-ture i n sublittoral soft sediments. Journal of Experimental Marine Biology and Ecology 91:247-262. Georgian, T . J . , Jr. and J .B . Wallace. 1981. A model of seston-capture by net-spinning caddisflies. Oikos 36:147-157. Gilbert , N . 1984. Control of fecundity i n Pieris rapae I. the problem. Journal of A n i m a l Ecology 53:581-588. Grafius, E . and N . H . Anderson. 1979. Population dynamics, bioenergetics, and role of Lepidostoma quercina Ross (Trichoptera: Lepidostomatidae) i n an Oregon woodland stream. Ecology 60:433-441. Guyer, C . 1988. Food supplementation i n a tropical mainland anole, Norops humilis: demographic effects. Ecology 69:350-361. Hairston, N . G . , F . E . Smith and L . B . Slobodkin. 1960. Community structure, pop-ulation control, and competition. American Naturalist 94:421-425. Hanson, B . J . , K . W . Cummins, J .R. Barnes and M . W . Carter. 1984. Leaf litter processing i n aquatic systems: a two variable model. Hydrobiologia 111:21-29. Hart , D . D . 1987. Experimental studies of exploitative competition i n a grazing stream insect. Oecologia (Berlin) 73:41-47. Hastings, A . 1984. Evolution i n a seasonal environment: simplicity lost? Evolution 38:350-358. Hemphil l , N . and S.D. Cooper. 1984. Differences i n the community structure of stream pools containing or lacking trout. Verhandlungen Internationale Vereinigung fur Theoretische und Angewandte Limnologie 22:1858-1861. Hildrew, A . G . and C R . Townsend. 1980. Aggregation, interference and foraging by larvae of Plectrocnemia conspersa (Trichoptera: Polycentropodidae). Animal Behaviour 28:553-560. 141 Hildrew, A . G . and C R . Townsend. 1987. Organization i n freshwater benthic com-munities, pp. 347-371 In: Gee J . H . R . and P.S. Giller (Eds.). Organization of Communities, Past and Present. Symposium of the Brit ish Ecological Society No. 27, Blackwell, Oxford. H i l l , W . R . and A . W . Knight . 1987. Experimental analysis of the grazing interaction between a mayfly and stream algae. Ecology 68:1955-1965. Holt , R . D . and B . P . Kotler . 1987. Short-term apparent competition. American Naturalist 130:412-430. Horn, H .S . and R . H . MacArthur . 1972. Competition among fugitive species i n a harlequin environment. Ecology 53:749-752. Howarth, R . W . and S .G. Fisher. 1976. Carbon, nitrogen, and phosphorus dynamics during leaf decay i n nutrient-enriched stream microecosystems. Freshwater Biology 6:221-228. Hunt , H . W . , D . C . Coleman, E . R . Ingham, R . E . Ingham, E . T . El l iott , J . C Moore, S.L. Rose, C P . P . Reid and C R . Morley. 1987. The detrital food web i n a shortgrass prairie. Biology and Fertility of Soils 3:57-68. H u r y n , A . D . and J .B . Wallace. 1987. Local geomorphology as a determinant of macrofaunal production i n a mountain stream. Ecology 68:1932-1942. Hynes, H . B . N . 1970. The Ecology of Running Waters. University of Toronto Press, 555 pp. Hynes, H . B . N . 1975. The stream and its valley. Verhandlungen Internationale Vereinigung fiir Theoretische und Angewandte Lirrmologie 19:1-15. Jewett, S .G. , Jr. 1959. The Stoneflies (Plecoptera) of the Pacific Northwest. Oregon State College, Corvallis, O R , 95 pp. Juliano, S .A. 1986. Food limitation of reproduction and survival for populations of Brachinus (Coleoptera: Carabidae). Ecology 67:1036-1045. Juliano, S .A. 1988. Chrysomelid beetles on water l i ly leaves: herbivore density, leaf survival, and herbivore maturation. Ecology 69:1294-1298. Kaushik, N . K . and H . B . N . Hynes. 1971. The fate of the dead leaves that fall into streams. Archiv fiir Hydrobiologie 68:465-515. Kerst, C D . and N . H . Anderson. 1975. The Plecoptera community of a small stream i n Oregon, U . S . A . Freshwater Biology 5:189-203. Kohler, S.L. 1985. Identification of stream drift mechanisms: an experimental and observational approach. Ecology 66:1749-1761. 142 Lambert i , G . A . , J . W . Feminella and V . H . Resli. 1987. Herbivory and intraspecific competition i n a stream caddisfly population. Oecologia (Berlin) 73:75-81. Lambert i , G . A . , S .V. Gregory, L . R . Ashkenas, R . C . W i l d m a n and A . D . Steinman. In Press. Influence of channel geomorphology on retention of dissolved and particulate matter i n a Cascade Mountain stream. In: Proceedings of the California Riparian Systems Conference. Levins, R . 1979. Coexistence i n a variable environment. American Naturalist 114:765-783. Lomnicki , A . 1988. Population Ecology of Individuals. Princeton, N J , 223 pp. Lush, D . L . and H . B . N . Hynes. 1973. The formation of particles i n freshwater leachates of dead leaves. Limnology and Oceanography 18:968-977. Mackay, R . J . and J . Kalff. 1969. Seasonal variation i n standing crop and species diversity of insect communities i n a small Quebec stream. Ecology 50:101-109. M a r t i n , M . M . , J . J . Kukar , J.S. M a r t i n , D . L . Lawson and R . W . Merrit t . 1981. D i -gestive enzymes of larvae of three species of caddisflies (Trichoptera). Insect Biochemistry 11:501-505. M a r t i n , M . M . , J.S. M a r t i n , J . J . Kukar and R . W . Merritt . 1980. The digestion of protein and carbohydrate by the stream detritivore, Tipula abdominalis (Diptera, Tipulidae). Oecologia (Berlin) 58:281-285. Mason, J . C . 1976. Response of underyearling Coho salmon to supplemental feeding i n a natural stream. Journal of Wildlife Management 40:775-788. Mattingly, R . L . 1987. Handling of coarse and fine particulate organic matter by the aquatic insects Paraleptophlebia gregalis and P. temporalis (Ephemeroptera: Leptophlebiidae). Freshwater Biology 18:255-265. May, R . M . 1981. Theoretical Ecology, Principles and Applications, Second Edit ion. Sinauer, Mass. M c A r t h u r , J . V . , J .R. Barnes, B . J . Hanson and L . G . Leff. 1988. Seasonal dynam-ics of leaf fitter breakdown i n a Utah alpine stream. Journal of the North American Benthological Society 7:44-50. M c A r t h u r , J . V . , L . G . Leff, D . A . Kovacic and J . Jaroscak. 1986. Green leaf decom-position i n coastal plain streams. Journal of Freshwater Ecology 3:553-558. McAulif fe , J .R . 1984. Resource depression by a stream herbivore: effects on distri-butions and abundances of other grazers. Oikos 42:327-333. McCauley, E . and W . W . Murdoch. 1987. Cyclic and stable populations: plankton as paradigm. American Naturalist 129:97-121. 143 McElravy , E . P . , G . A . Lamberti and V . H . Rest . 1989. Year-to-year variation i n the aquatic macroinvertebrate fauna of a northern California stream. Journal of the North American Benthological Society 8:51-63. McQueen, D . J . and J .R. Post. 1988. Cascading trophic interactions: uncouphng at the zooplankton-phytoplankton l ink. Hydrobiologia 159:277-296. McQueen, D . J . , J .R . Post and E . L . Mi l l s . 1986. Trophic relationships i n freshwa-ter pelagic ecosystems. Canadian Journal of Fisheries and Aquatic Sciences 43:1571-1581. Meli l lo , J . M . , R . J . Naiman, J . D . Aber and K . N . Eshleman. 1983. The influence of substrate quality and stream size on wood decomposition dynamics. Oecolo-gia (Berlin) 58:281-285. Merr i t t , R . W . and K . W . Cummins. 1984. A n Introduction to the Aquatic Insects of North America, 2 n d Edit ion. Kendall-Hunt, Iowa, 722 pp. Minshal l , G . W . 1967. Role of allochthonous detritus i n the trophic structure of a woodland springbrook community. Ecology 48:139-149. Minshal l , G . W . , R . C . Petersen, K . W . Cummins, T . L . Bott , J .R. Sedell, C E . Gush-ing and R . L . Vannote. 1983. Interbiome comparison of stream ecosystem dynamics. Ecological Monographs 53:1-25. Minshal l , G . W . and J . N . Minshall . 1977. Microdistribution of benthic macroinver-tebrates in a Rocky Mountain (U.S.A.) stream. Hydrobiologia 55:231-249. Molles, M . C . , J r . and R . D . Pietruszka. 1983. Mechanisms of prey selection by predaceous stoneflies: roles of prey morphology, prey behavior and predator hunger. Oecologia (Berlin) 57:25-31. Mulhol land, P . J . , J . W . Elwood, J . D . Newbold and L A . Ferren. 1985. Effect of a leaf-shredding invertebrate on organic matter dynamics and phosphorus spiralling i n heterotrophic laboratory streams. Oecologia (Berlin) 66:199-206. Mundie , J . H . , S . M . McKinne l l and R . E . Traber. 1983. Responses of stream zooben-thos to enrichment of gravel substrates with cereal grain and soybean. Cana-dian Journal of Fisheries and Aquatic Sciences 40:1702-1712. Mundie , J . H . , D . E . Mounce and L . E . Smith. 1973. Observations on the response of zoobenthos to additions of hay, willow leaves, alder leaves and cereal grain to stream substrates. Fisheries Research Board of Canada Technical Report No. 387. Murphy, M . L . , C P . Hawkins and N . H . Anderson. 1981. Effects of canopy modi-fication and accumulated sediment on stream communities. Transactions of the American Fisheries Society 110:469-478. 144 Naiman, R . J . 1983. The influence of stream size on the food quality of seston. Canadian Journal of Zoology 61:1995-2010. Naiman, R . J . , J . M . Meli l lo , M . A . Lock, T . E . Ford and S.R. Reice. 1987. Longitudi-nal patterns of ecosystem processes and community structure i n a subarctic river continuum. Ecology 68:1139-1156. Naiman, R . J . and J .R. Sedell. 1979. Benthic organic matter as a function of stream order i n Oregon. Archiv fiir Hydrobiologie 87:404-422. Neaves, P.I. 1978. Litterfall , export, decomposition, and retention i n Carnation Creek, Vancouver Island. Fisheries and Marine Service Technical Report No. 809:1-43. NeiU, W . E . 1975. Experimental studies of microcrustacean competition, community composition and efficiency of resource utilization. Ecology 56:809-826. Nei l l , W . E . 1988. Complex interactions in oligotrophic lake food webs: responses to nutrient enrichment, pp.31-44 In: Carpenter, S.R. (Ed.). Complex Inter-actions i n Lake Communities. Springer-Verlag, N Y . NeiU, W . E . and A . Peacock. 1980. Breaking the bottleneck: interactions of in -vertebrate predators and nutrients i n oligotrophic lakes, pp. 715-725 i n : Kerfoot, W . C . (ed.), Evolution and Ecology of Zooplankton Communities. University Press of New England, Hanover, N H . Nelson, D . J . and D . C . Scott. 1962. Role of detritus in the productivity of a rock-outcrop community i n a piedmont stream. Limnology and Oceanography 7:396-413. Nicholson, A . J . 1954. A n outline of the dynamics of animal populations. Australian Journal of Zoology 2:9-65. N i s b e t , R . M . and W . S . C . Gurney. 1976. Populations dynamics i n a periodically varying environment. Journal of Theoretical Biology 56:459-475. Northcote, T . G . 1988. Fish in the structure and function of freshwater ecosystems: a "top-down" view. Canadian Journal of Fisheries and Aquatic Sciences 45:361-379. Oberndorfer, R . Y . , J . V . M c A r t h u r , J .R . Barnes and J . Dixon. 1984 The effect of invertebrate predators on leaf litter processing i n an alpine stream. Ecology 65:1325-1331. Oksanen, L . 1988. Ecosystem organization: mutualism and cybernetics or plain Darwinian struggle for existence? American Naturalist 131:424-444. Oksanen, L . , S .D. Fretwell, J . Arruda and P. Niemela. 1981. Exploitation ecosys-tems i n gradients of primary productivity. American Naturalist 118:240-261. 145 Otto, C . 1974. Growth and energetics i n a larval population of Potamophylax cingu-latus (Steph.) (Trichoptera) i n a south Swedish stream. Journal of Animal Ecology 43:339-361. Pearson, T . H . and R. Rosenberg. 1978. Macrobenthic succession i n relation to organic enrichment and pollution of the marine environment. Oceanography and Marine Biology, A n Annual Review 16:229-311. Pearson, T . H . and R . Rosenberg. 1987. Feast and famine: structuring factors i n marine benthic communities, pp. 373-395 In: Gee, J . H . R . and P.S. Giller [Eds.]. Organization of Communities, Past and Present. Brit ish Ecological Society Special Publication no. 27, Blackwell, Boston. Peckarsky, B . L . and M . A . Penton. 1985. Is predaceous stonefly behavior affected by competition? Ecology 66:1718-1728. Pennak, R . W . 1978. Fresh-Water Invertebrates of the United States, 2nd Edit ion. Wiley - Interscience, Toronto. Petersen, R . C , Jr. 1987. Seston quality as a factor influencing Trichopteran popu-lations, pp. 287-292 In: Bournaud, M and H . Tachet [Eds.]. Proceedings of the F i f th International Symposium on Trichoptera. D r . W.Junk Publishers, Boston. Petersen, R . C , Jr. and K . W . Cummins. 1974. Leaf processing i n a woodland stream. Freshwater Biology 4:343-368. Petersen, R . C , Jr., K . W . Cummins and G . M . Ward. 1989. Microbial and animal processing of detritus i n a woodland stream. Ecological Monographs 59:21-39. P i m m , S.L. 1982. Food Webs. Chapman and Hal l , New York. Power, M . E . , W . J . Mathews and A . J . Stewart. 1985. Grazing minnows, piscivorous bass, and stream algae: dynamics of a strong interaction. Ecology 66:1448-1456. Rabeni, C F . and G . W . Minshall . 1977. Factors affecting microdistribution of stream benthic insects. Oikos 29:33-43. Radford, D.S. and R. Hartland-Rowe. 1971. The life cycles of some stream insects (Ephemeroptera, Plecoptera) i n Alberta. Canadian Entomologist 103:609-617. Reice, S.R. 1974. Environmental patchiness and the breakdown of leaf Utter i n a woodland stream. Ecology 55:1271-1282. Reice, S.R. 1980. The role of substratum in benthic macroinvertebrate microdistri-bution and litter decomposition i n a woodland stream. Ecology 61:580-590. 146 Reice,S.R. and R . L . Edwards. 1986. The effect of vertebrate predation on lotic macroinvertebrate communities i n Quebec, Canada. Canadian Journal of Zoology 64:1930-1936. Rietsma, C.S. , I. Valiela and R. Buchsbaum. 1988. Detrital chemistry, growth, and food choice i n the salt-marsh snail (Melampus bidentatus). Ecology 69:261-266. Ricker, W . E . 1943. Stoneflies of southwestern Brit ish Columbia. Indiana University Publications Science Series 12:1-145. Rottenberry, J . A . and J . A . Wiens. 1985. Statistical power analysis and community-wide patterns. American Naturalist 125:164-168. Roughgarden, J . 1986. A comparison of food-limited and space-limited animal competition communities, pp. 492-516. i n : Diamond, J . and T . J . Case (eds.) Community Ecology. Harper and Row, N Y . SAS Institute Inc. 1985. S A S / S T A T Guide for Personal Computers, Version 6 Edit ion. Cary, N C , U S A . 378pp. Sakaguchi, D . K . 1978. Two Zapada species i n a Rocky Mountain stream: life his-tories and feeding habits (Nemouridae: Plecoptera). M.Sc. Thesis, Brigham Young University, Utah. Schoener, T . W . 1982. The controversy over interspecific competition. American Scientist 70:586-595. Schoener, T . W . 1983. Field experiments on interspecific competition. American Naturalist 122:240-285. Sedell, J .R. , F . J . Triska and N.S. Triska. 1975. The processing of conifer and hard-wood leaves i n two coniferous forest streams: I. Weight loss and associated invertebrates. Verhandlungen Internationale Vereinigung fiir Theoretische und Angewandte Limnologie 19:1617-1627. Sheldon, A . L . and S .G. Jewett, Jr. 1967. Stonefiy emergence i n a Sierra Nevada stream (Plecoptera). Pan-Pacific Entomologist 43:1-8. Shepard, R . B . and G . W . Minshall . 1984. Role of benthic insect feces i n a Rocky Mountain stream: fecal production and support of consumer growth. Hol-arctic Ecology 7:119-127. Short, R . A . , S.P. Canton and J . V . Ward. 1980. Detrital processing and associated macroinvertebrates in a Colorada mountain stream. Ecology 61:727-732. Short, R . A . and J . V . Ward. 1981. Trophic ecology of three winter stoneflies (Ple-coptera). American Midland Naturalist 105:341-347. 147 Short, R . A . and P . E . Masl in . 1977. Processing of leaf litter by a stream detritivore: effect on nutrient availability to collectors. Ecology 58:935-938. Short, R . A . and S.L. Smith. 1989. Seasonal comparison of leaf processing i n a Texas stream. American Midland Naturalist 121:219-224. Shortreed, K . S . and J . G . Stockner. 1983. Periphyton biomass and species composi-tion i n a coastal rainforest stream i n Brit ish Columbia: effects of environmen-tal changes caused by logging. Canadian Journal of Fisheries and Aquatic Sciences 40:1887-1895. Shiozawa, D . K . 1983. Density independence versus density dependence i n streams, pp. 55-77 i n : Barnes, J .R. and G . W . Minshall [Eds.]. Stream Ecology: application and testing of general ecological theory. Plenum Press, N Y . Sibly, R . and P. Calow. 1987. Ecological compensation - a complication for testing life-history theory. Journal of Theoretical Biology 125:177-186. Sinclair, A . R . E . In press. The regulation of animal populations. In: Symposium of the Bri t ish Ecological Society, Blackwell, Oxford. Soluk, D . A . and N . C . Collins. 1988. Synergistic interactions between fish and stoneflies: facilitation and interference among stream predators. Oikos 52:94-100. Speaker, R. , K . Moore and S. Gregory. 1984. Analysis of the process of reten-tion of organic matter i n stream ecosystems. Verhandlungen Internationale Vereinigung fur Theoretische und Angewandte Limnologie 22:1835-1841. Stanford, J . A . and J . V . Ward. 1983. Insect species diversity as a function of environmental variability and disturbance i n stream systems, pp.265-278 In: Barnes, J .R. and G . W . Minshall [Eds.]. Stream Ecology: application and testing of general ecological theory. Plenum Press, N Y . Stout, R . J . and W . H . Taft. 1985. Growth patterns of a chironomid shredder on fresh and senescent tag alder leaves in two Michigan streams. Journal of Freshwater Ecology 3:147-153. Stout, R . J . , W . H . Taft and R . W . Merritt . 1985 Patterns of macroinvertebrate colonization on fresh and scenescent alder leaves in two Michigan streams. Freshwater Biology 15:573-580. Street, M . and G . Titmus. 1982. A field experiment on the value of allochthonous straw as food and substratum for lake macro-invertebrates. Freshwater Biol-ogy 12:403-410. Sweeney, B . W . and R . L . Vannote. 1986. Growth and production of a stream stonefly: influences of diet and temperature. Ecology 67:1396-1410. 148 Sweeney, B . W . , R . L . Vannote and P . J . Dodds. 1986. Effects of temperature and food quality on growth and development of a mayfly, Leptophlebia intermedia. Canadian Journal of Fisheries and Aquatic Sciences 43:12-18. Swift, M . J . , O . W . Heal and J . M . Anderson. 1979. Decomposition i n Terrestrial Ecosystems. Blackwell , London. T i l m a n , D . 1982. Resource Competition and Community Structure. Princeton University Press, Princeton, N J . Toft, C . A . and P . J . Shea. 1983. Detecting community-wide patterns: estimating power strengthens statistical inference. American Naturalist 122:618-625. Triska, F . J . and K . Cromack, Jr. 1980. The role of wood debris i n forests and streams. In: Waring, R . H . (Ed.) Forests: Fresh Perspectives from Ecosystem Analysis. Oregon State University Press, O R . Triska, F . J . , J .R . Sedell and B . Buckley. 1975. The processing of conifer and hard-wood leaves i n two coniferous forest streams: II. Biochemical and nutrient changes. Verhandlungen Internationale Vereinigung fur Theoretische und Angewandte Limnologie 19:1628-1639. Triska, F . J . , J .R . Sedell, K . Cromack,Jr. , S .V. Gregory and F . M . McCorison. 1984. Nitrogen budget for a small coniferous forest stream. Ecological Monographs 54:119-140. Triska, F . J . , J .R . Sedell and S.V. Gregory. 1982. Coniferous forest streams, pp. 292-332 In: Edmonds, R . L . (Ed.). Analysis of Coniferous Forest Streams i n the Western United States. Hutchinson Ross, P A . Vallett , H . M . and J . A . Stanford. 1987. Food quality and hydropsychid caddisfly density i n a lake outlet stream i n Glacier National Park, Montana, U S A . Canadian Journal of Fisheries and Aquatic Sciences 44:77-82. Vannote, R . L . , G . W . Minshal l , K . W . Cummins, J .R. Sedell and C . E . Cushing. 1980. The river continuum concept. Canadian Journal of Fisheries and Aquatic Sciences 37:370-377. Walde, S.J. and R . W . Davies. 1984. The effect of intraspecific interference on Ko-gotus nonus (Plecoptera) foraging behaviour. Canadian Journal of Zoology 62:2221-2226. Wallace, J .B . , J .R . Webster and T . F . Cuffney. 1982. Stream detritus dynamics: regulation by invertebrate consumers^. Oecologia (Berlin) 53:197-200. Wallace, J .B . , W . R . Woodall and F . F . Sherberger. 1970. Breakdown of leaves by feeding of Peltoperla maria nymphs (Plecoptera: Peltoperlidae). Annals of the Entomological Society of America 63:562-567. 149 Walters, C . J . , E . Krause, W . E . Neil l and T . G . Northcote. 1987. Equil ibrium models for seasonal dynamics of plankton biomass i n four oligotrophic lakes. Cana-dian Journal of Fisheries and Aquatic Sciences 44:1002-1017. W a r d , G . M . and N . G . Aumen. 1986. Woody debris as a source of fine particulate organic matter i n coniferous forest stream ecosystems. Canadian Journal of Fisheries and Aquatic Sciences 43:1635-1642. W a r d , G . M . and K . W . Cummins. 1979. Effects of food quality on growth of a stream detritivore, Paratendipes albimanus (Meigan) (Diptera: Chironomi-dae). Ecology 60:57-64. Warren, C . E . , J . H . Wales, G . E . Davis and P. Doudoroff. 1964. Trout production i n an experimental stream enriched with sucrose. Journal of Wildlife Manage-ment 28:617-660. Waters, T . F . 1972. The drift of stream insects. Annual Review of Entomology 17:253-272. Webster, J .R . 1983. The role of benthic macroinvertebrates i n detritus dynamics of streams: a computer simulation. Ecological Monographs 53:383-404. Webster, J .R . and E . F . Benfield. 1986. Vascular plant breakdown i n freshwater ecosystems. Annual Review of Ecology and Systematics 17:567-594. Webster, J .R. , M . E . Gurtz , J . J . Hains, J . L . Meyer, W . T . Swank, J .B . Waide and J . B . Wallace. 1983. Stability of stream ecosystems, pp, 355-395 In: Barnes, J .R . and G . W . Minshall [Eds.]. Stream Ecology: application and testing of general ecological theory. P lenum Press, N Y . Webster, J .R . and J .B . Waide. 1982. Effects of forest clearcutting on leaf breakdown i n a southern Appalachian stream. Freshwater Biology 12:331-344. White , T . C . R . 1978. The importance of a relative shortage of food i n animal ecology. Oecologia (Berlin) 33:71-86. Wil l iams, D . D . , J . H . Mundie and D . E . Mounce. 1977. Some aspects of benthic production i n a salmonid rearing channel. Journal of the Fisheries Research Board of Canada 34:2133-2141. Winterbourn, M . J . 1971. The life histories and trophic relationships of the Tr i -choptera of Marion Lake, Brit ish Columbia. Canadian Journal of Zoology 49:623-635. 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. 150 Wolda, H . 1988. Insect seasonality: why? Annual Review of Ecology and System-atics 19:1-18. Yodzis, P. 1988. The indeterminacy of ecological interactions as perceived through perturbation experiments. Ecology 69:508-515. 151 A p p e n d i x 1 — Preliminary list of detritivore taxa from Mayfly Creek, Brit ish Columbia. P L E C O P T E R A Nemouridae Zapada cinctipes (Banks) Zapada haysi (Bicker) Zapada frigida (Claassen) Malenka californica (Claassen) Malenka cornuta (Claassen) Soyedina producta (Claassen) Vis oka cataractae (Neave) Podmosta sp. Leuctridae Despaxia augusta (Banks) Moselia infuscata (Claassen) Capniidae Capnia nana Claassen Capnia sp. Pteronarcyidae Pteronarcys californica Newport Peltoperlidae Yoraperla mariana (Bicker) T R I C H O P T E R A Lepidostomatidae Lepidostoma roafi (Milne) Lepidostoma unicolor (Banks) Lepidostoma cascadense (Milne) Limnephilidae Onocosmoecus unicolor (Banks) Dicosmoecus atripes (Hagen) Ecclisocosmoecus scylla (Milne) Ecclisomyia conspersa Banks Psychoglypha species D I P T E R A Chironomidae Brillia retifinis Saether 152 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0106745/manifest

Comment

Related Items