Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Ecology and condition of the ground beetle Scaphinotus angusticollis and distribution of its prey in… Lavallee, Susanne L. 2006

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

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-ubc_2006-200117.pdf [ 5.4MB ]
Metadata
JSON: 831-1.0075024.json
JSON-LD: 831-1.0075024-ld.json
RDF/XML (Pretty): 831-1.0075024-rdf.xml
RDF/JSON: 831-1.0075024-rdf.json
Turtle: 831-1.0075024-turtle.txt
N-Triples: 831-1.0075024-rdf-ntriples.txt
Original Record: 831-1.0075024-source.json
Full Text
831-1.0075024-fulltext.txt
Citation
831-1.0075024.ris

Full Text

Ecology and Condit ion of the Ground Beet le Scaphinotus angusticollis and Distribution of its Prey in Paci f ic Northwest Ripar ian Forests by S u s a n n e L. Laval lee B . S c (University of British Columbia) , 1994 M . S c (University of British Columbia , Zoology), 1999 A T H E S I S S U B M I T T E D IN 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 In 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 (Forestry) T H E U N I V E R S I T Y O F BRIT ISH C O L U M B I A September 2006 © S u s a n n e L. Laval lee, 2006 Abstract I studied the population ecology of the f l ightless, forest-dwell ing carabid beetle Scaphinotus angusticollis F ischer V o n Wa ldhe im (O. Co leoptera F. Carab idae) and severa l aspec ts of its body condit ion for their assoc ia t ions with forest harvest ing. Compar i son of population est imates revealed that catch-per-unit-effort est imates were not significantly different from more detai led ana lyses . In two years of trapping, S . angusticollis population densit ies were found to be significantly lower in c learcuts, a s compared to 30 m riparian reserves and uncut forest, suggest ing that riparian buffers provide adequate habitat to maintain populat ions of this terrestrial insect. Movement of S. angusticollis differed in the three habitats studied between years and treatments, with the greatest movement occurr ing in 30 m buffers in one year and in control si tes the next. C learcuts had the lowest amount of movement in both years. O n e of the known prey spec ies for S . angusticollis, snai ls < 2 c m in diameter, were more abundant in clearcut habitat, with Ancotrema hybridum the most abundant spec ies . Canon ica l cor respondence analys is suggests that only A. hybridum were positively correlated with plant cover, and that other spec ies abundances may rely on coarse , downed wood as cover. Thus only some spec ies of snai ls are assoc ia ted with recently logged a reas . The internal body condit ions of S . angusticollis living in clearcut, 30 m buffer and uncut sites showed that gut ful lness w a s significantly correlated with energy storage (fat body) and habitat types (clearcut, 30 m buffer and uncut). Th is study demonstrates how some forest insects, particularly ground beet les, are affected by harvesting and that current management pract ices on the stand sca le can mitigate some of the negative impacts of logging. Table of Contents Abstract ii Tab le of Contents iii List of Tab les v List of F igures vi Acknowledgements '. viii Chapter O n e Introduction 1 Questions 7 Study Organism 8 Literature Cited 9 Chapter Two Evaluat ion of two populat ion est imate methods for ground beetle (O. Co leoptera , F. Carab idae) populat ions 16 Introduction 16 Methods '. :....18 Results : 22 Discussion 24 Literature Cited 28 Figures 31 Chapter Three Relat ive abundance and movement of the carabid beetle Scaphinotus angusticollis in coniferous riparian forests of the Paci f ic Northwest 33 Introduction 33 Methods 37 Results 42 Discussion 44 Literature Cited .'. 49 Tables 54 Figures 56 Chapter Four Snai l (Gast ropoda, Pulmonata) abundance and diversity in three forest habitats of Paci f ic Northwest coni ferous forests (Canada) 59 Introduction 59 Methods 61 Results : 63 Discussion .• 64 Literature Cited 68 Tables 70 Figures 71 Chapter Five Forest harvest ing and the assoc ia ted body condit ions of the carabid beetle Scaphinotus angusticollis 72 Introduction ; 72 Materials and Methods 75 Results 79 Discussion 84 Literature Cited 88 Tables 92 Figures 96 Chapter S ix Conc lus ion , 99 Findings of this study 99 Use of riparian reserves by forest-preferring species 101 i n Use of indicator taxa 102 Contributions to insect ecology and conservation... 103 Literature Cited 106 Append ix I Food preference in Scaphinotus angusticollis (F ischer V o n Waldhe im) 109 Introduction 109 Methods : 109 Results : 111 Discussion ; 112 Literature Cited 114 iv List of Tables Tab le 3.1 Ana lys is of var iance of catch per unit effort ( C P U E ) for Scaphinotus angusticollis in forest, 30 m buffer, and clearcut si tes. Main effects include habitat type (H), Si te (H), month, the interaction of month and habitat and year. (Model r2 = 0.49, df = 16) ,: .'. . 54 Tab le 3.2 Ana lys is of var iance of catch per unit effort population est imates for Scaphinotus angusticollis in forest, 30 m reserve, and clearcut si tes at varying d is tances from the stream. Main effects include habitat type (H), the nested term of row within habitat type, month, and the interaction of month and habitat. (Model r2 = 0,72, df = 41) 54 Tab le 3.3 Ana lys is of var iance for d is tances moved by Scaphinotus angusticollis in clearcut, 30 m reserve and control sites at d is tances less than 30 m from the st ream (close) and greater than 30 m from the stream (far). Main effects include habitat (H), site nested within habitat, and "c lose or far" nested within habitat. (Model r2 = 0.94, df = 10) 55 Tab le 3.4 Total number of recaptures of S . angusticollis made in control, 30 m buffer and clearcut sites in 2000 and 2001 . Si te names are provided under each habitat type •. 55 Tab le 3.5 Ana lys is of var iance of d is tances moved by Scaphinotus angusticollis in clearcut, 30 m reserve and control sites. Main effect is habitat only. (Model r2 = 0.07, df = 2) 55 Tab le 4.1 Canon ica l cor respondence analys is results compar ing spec ies occur rences with habitat features (plant cover, coarse woody debr is, and fine woody debris).. . 70 Tab le 5.1 Summary of trapping dates and sites, with total number of Scaphinotus angusticollis d issected per t ime period and site. Ze roes indicate that no individuals were caught 92 Tab le 5.2 Ana lys is of var iance of gut ful lness for Scaphinotus angusticollis in forest, 30 m reserve, and clearcut s i tes. Main effects include habitat type (H), site nested within habitat, date and the interaction of date and habitat type. (Model r2 = 0.38, df = 14). ; ........... 93 Tab le 5.3 Ana lys is of var iance of fat body for Scaphinotus angusticollis in forest, 30 m reserve, and clearcut sites. Main effects include habitat type (H), date nested within habitat type, and site nested within habitat type. (Model r = 0.45, df = 14) 93 Tab le 5.4 Ordinal logistic fit for rank of sex development for Scaphinotus angusticollis in forest, 30 m reserve and clearcut sites. Main effects include habitat type (H), date nested within habitat type, and site nested within habitat type. (Model r2 = 0.43, df = 14) 93 Tab le 5.5 Ana lys is of var iance results of water content of Scaph inotus angusticol l is in forest, 30 m reserve and clearcut sites. Main effects include habitat type (H), site nested within habitat, date and the interaction of date and habitat. (Model r2 = 0.74, df = 14) 94 Tab le 5.6 Resu l ts for ana lyses of covar iance tests of weight/pronotum ratio with gut fu l lness, fat body, water content and sex Scaphinotus angusticollis 95 List of Figures Figure 2.1 Compar i son of population est imates for the adult carabid Scaphinotus angusticollis sampled in 2000 and 2001 in control, 30 m buffer and clearcut habitats. Est imates were made from pitfall trapping and individual marking of beet les. A ) Schnabe l -Schumache r and C P U E est imates; B) Corrected Schnabe l -S c h u m a c h e r and C P U E est imates; and C) Corrected Schnabe l -Schumache r and Schnabe l -Schumache r est imates. Dashed line represents 1:1 ratio 31 Figure 2.2 Conco rdance correlation results, compar ing catch-per-unit-effort ( C P U E ) and Schnabe l -Shumache r (SS) population est imates (circles) and catch-per-unit-effort ( C P U E ) and corrected Schnabe l -Schumache r ( cSS) population est imates (squares) in each of the three habitats: control, 30 m buffer and clearcut si tes 32 Figure 2.3 Concordance correlation analys is results for all Scaphinotus angusticollis populat ion est imates, pooled by year 32 Figure 3.1 Catch-per-unit-effort ( C P U E ) population est imates (not including any recaptures) of S . angusticollis for control (circle), 30 m buffer (triangle) and clearcut (square) si tes in 2000, 2001 and 2002. Error bars indicate standard error. N=3 for all trapping dates except N=2 September /October 2000 and 2002. Aster isk (*) indicates significant difference (p = 0.03) 56 Figure 3.2 Distribution of mean numbers of S . angusticollis caught per trap night in control, 30 m buffer and clearcut habitats in 2000 (A) and 2001 (B) at varying d is tances from the st ream. Dotted line indicates forest boundary in 30 m buffer si tes. Error bars represent standard error. (N (number of trapping sess ions) = 11 (control 2000) 12 (30 m buffer 2000) 11 (clearcut 2000) and N = 14 (all habitats 2001)) 57 Figure 3.3 M e a n distance moved overnight by S. angusticollis in control (circle), 30 m buffer (triangle), and clearcut (square) habitat in 2000 (black) and 2001 (grey). Error bars represent standard error. S e e Table 3.3 for N of e a c h point 58 Figure 3.4 M e a n movement vectors (In) for d is tances travelled overnight by S . angusticollis in control, 30 m buffer and clearcut sites in 2000 and 2001 (pooled). Error bars represent standard error. N= 7 (clearcut), 26 (30 m buffer), and 15 (control), not weighted by site within habitat 58 Figure 4.1 M e a n number of snai ls found in control, 30m buffer and clearcut habitats a long 45 m transects in 2002. N=3 for each habitat type. Error bars indicate standard error 71 Figure 4.2 M e a n Shannon -Weave r Diversity Index (FT) va lues for control, 30m buffer and clearcut site snai l diversity in 2003. N=3 for all habitat types, error bars indicate standard error 71 Figure 5.1 M e a n total gut ful lness (A) and fat body (B) in beet les from control, 30 m buffer and clearcut sites in 2002. N= 2-3, depending on data point; error bars indicate standard error. Aster isk (*) indicates significant difference. 96 Figure 5.2 M e a n binary reproductive rank of beet les in control, 30m buffer and clearcut habitats. N = 2-3, depending on data point; error bars indicate standard error 97 Figure 5.3 Water content of beet les controlled for body s ize (wet - dry weight/ wet weight) from control, 30 m buffer and clearcut habitats in 2002. N = 2-3, depending on data point; error bars indicate standard error. 97 V I Figure 5.4 M e a n ln(weight+1)/pronotum ratios (g/cm) for beet les caught in control, 30 m buffer and clearcut habitats in 2000-2002. N = 2-3, depending on data point; error bars indicate standard error. Months correspond directly to the ca lendar months (i.e. 5=May, 6=June, etc.) .98 Append ix Figure 1.1 Consumpt ion of foods by Scaphinotus angusticollis 115 v i i Acknowledgements My s incere thanks go out to many people who helped me over the years. My supervisor, John R ichardson , let me explore many crazy ideas and didn't say "I told you so" too often. My committee, Peter A r c e s e , Judy Myers, Jam ie Smith, and John Mc lean , gave insightful comments at the right moments and chal lenged me when I needed it. My supervisor for my M . S c , Geoff Scudder , has been an inspiration to me for work in insect conservat ion and my love of field studies. My field assistants Jay Haddrel l , Karen V ickers , Mike Nivens, and Janneke Lade were patient and hard workers. M y many lab mates over the years l istened to my rants and offered some sanity in the worst moments. My co l leagues in teaching Caro l Pol lock, Kathy Nomme and Lynn Norman helped me survive first year biology many t imes over. My fr iends Jennifer Heron, Darren B e n s o n , Karen Needham, Jeff Lemieux, Tamara Sib ley, Marja Mackenz ie , Emi ly G o n z a l e s , Katie Aitkin, Kerri Bates , J e s s i c a Meyer -Rachner , Car in Bondar , Laurie Miller, Kather ine Maxcy, J im Herbers, C o n a n Phe lan , R o s s Thompson , and Tat iana Lee have all contributed to my work and my life in different and valuable ways . This thesis is dedicated to my daughter, Mar ie, who cont inues to inspire me to be a better teacher and my parents, J e a n and E d , who have always inspired me to be my best. v n i Chapter One Introduction Insects are vital to all terrestrial food webs as both predators (Symondson et al. 2002, Paetzo ld and Tockner 2005) and prey. In addit ion, they provide energy l inkages across terrestrial/aquatic boundar ies (Hering 1998, Nakano et al. 1999, Paetzo ld etal. 2005). S o m e guilds of insects are pivotal to arthropod and bacteria, nematode and eukaryote diversity (Boulton and Ambe rman 2006); others may be cons idered to be ecosys tem engineers by extensively modifying the physical structure of their physical habitat (Moore 2006). In forests, saproxyl ic insects accelerate the process of decay in both standing and downed wood (Grove 2002) and provide s o m e level of pest insect control through predation (Symondson et al. 2002). Despi te the many vital roles that insects play, insect conservat ion is relatively neglected (Dunn 2005), especia l ly in forest environments. O n e notable except ion are the saproxyl ic insects, which have become a focus for conservat ion efforts in Fennoscand ian forestry (Hyvar inen et al. 2005 , J o n s s o n et al. 2006). Carab id beet les have a lso become the focus of conservat ion researchers in J a p a n (Osawa et al. 2005), Europe (Kotze and O 'Hara 2003), Fennoscand ia (Koivula and Niemela 2003), Tasman ia (Grove and Yax ley 2005) and C a n a d a (K l imaszewsk i et al. 2005), but are rarely the target for conservat ion policy or other efforts. Harvest ing fragments forests and reduces the amount of forested habitat avai lable (Po lasky et al. 2005). Whi le there may be some transient, positive effects of habitat fragmentation on some insect populat ions (Grez et al. 2004), spec ies of insects that rely on patchily-distributed resources may decl ine after forest harvesting (Col l inge et al. 2001). In forest harvest ing, planning may include the maintenance of structural 1 features intended to provide habitat for forest spec ies , but few monitor the use of these features by wildlife (Work et al. 2003). Forest managers must plan on different spatial sca les for the retention of forested habitat (Swanson and Franklin 1991). The inclusion of forest reserves poses an additional chal lenge to the already difficult task of balancing the amount of intact forest "core" habitat to leave in areas scheduled for logging (Potvin et al. 2005). Smal le r -sca le planning may include wildlife tree patches, riparian reserves, and cutblock edge des ign (Swanson and Franklin 1991), each with different intended ou tcomes. Al l remaining a reas are ultimately des igned to provide habitat to spec ies that are typically dependent on forest-core habitat (forest-preferring spec ies) or are rarely found outside of forest-core habitat (forest-dependent spec ies) (Harris and P i m m 2004). F rom a conservat ion perspect ive, riparian a reas may also harbour different spec ies , not just more of them; in order to conserve a host of spec ies , riparian buffers must be included in landscape planning (Sabo et al. 2005). Ripar ian buffers have some benefits and drawbacks. Despi te their c o m m o n use in the Paci f ic Northwest (Moore 2005), there are increased costs to forest harvesting and the retention of forest set -as ides may be constrained by soc ioeconomic aspec ts of planning (Po lasky et al. 2005) and not biological requirements of spec ies that use them. However, retention of riparian forests may also protect old-growth t imber (Lee and Barker 2005). Headwater s t reams pose further chal lenges to forest managers . Whi le forest planning in British Co lumb ia does not require any protection for headwater s t reams unless they have fish or are a part of a drinking water supply (Moore 2005), p rocesses 2 in headwater s t reams are vital to the forests that surround them (Richardson et al. 2005) and the current division into f ish-bearing (protected) versus non-f ish-bearing (not protected) s t reams fails to recognize the contributions that smal ler s t reams make to waterways (Cummins and Wi l zbach 2005, Wipfl i 2005). Paci f ic Northwest forests are unique in that they tend to have little transition from riparian vegetat ion to upslope forest, and very few alluvial deposi ts (R ichardson et al. 2005). There are a lso a large number of smal l s t reams in mes ic forests (Richardson et al. 2005), but lack of understanding about the ecological functions of smal l s t reams and their riparian z o n e s has contributed to the uncertainty over how much to leave in buffers a long the s ides of s t reams (Moore and Richardson 2003). In addit ion to these cha l lenges in the traditional application of riparian buffers, forest management planning in British Co lumb ia has taken on a unique view: riparian corridors are a lso intended to serve as refuge habitat for forest-preferring spec ies (Anonymous 1995). Current management approaches that extend the benefits of riparian buffers to terrestrial communit ies need c loser examinat ion in a number of ways . T h e current management paradigm for riparian buffers is to cons ider width (Shir ley 2006) and not the matrix that surrounds it, which may not be sufficient for wider-ranging, terrestrial spec ies (Roderwald and Bakermans 2006). Most work on spec ies not dependent on riparian habitat has been on vertebrates (Richardson et al. 2005) but the cal l for studying the use of riparian habitat by forest-preferring spec ies needs to be answered for a broad diversity of organisms (Hylander et al. 2004). In order to maximize the probability of detecting the impacts of forest management , practitioners should use a range of spec ies from different functional 3 groups that use different features of forests (Taylor and Doran 2001). O n e of the most common approaches in monitoring is to use indicator spec ies , which are intended to show the effects of habitat disturbance on the biota or shifts in a host of spec ies (Caro and O'Doherty 1999). However , biodiversity management should not just amount to indicator spec ies management , as the relat ionships may change between indicators and the condit ions or spec ies they indicate (Simberloff 1999). Invertebrates ought to be a vital part of any monitoring programme; without adequate representat ion for insects, monitoring would not adequately evaluate the management impacts on the biota (Taylor and Doran 2001). Insects have an intermediate length of l i fecycle, ensur ing both sensitivity and stability of responses to environmental condit ions (Hodkinson and J a c k s o n 2005). Their many vital roles in forested ecosys tems a lso justify their inclusion in monitoring. Of the invertebrate fauna, carabid beet les (F. Carab idae) have been found by severa l studies to be very good bioindicators of habitat d isturbance (Oliver et al. 2000 , Rain io and Niemela 2002), making this taxon an excel lent starting point for understanding how forest insect ecology is affected by harvest ing. With speci f ic application to riparian a reas , carabid beet les have been used a s indicators for riverine management , to evaluate stream bank condit ions (Van Looy et al. 2005). There are several practical and theoretical advantages to using carabid beet les as indicators of habitat d isturbance in the forested setting. There is a wealth of information on the habitat preferences (Lindroth 1961, Thie le 1977, Lensk i 1982) of carabid beetles, but few studies have examined the functional relat ionships between these insects and their environment. In riparian ecosys tems , carabid beet les link 4 aquat ic and terrestrial p rocesses through predation, receiving significant amounts of food from emerging st ream insects (Hering 1998) and influencing the number of insects emerging from st reams (Paetzold and Tockner 2005). There is some concern over the conservat ion of carabid beetles, a s dec l ines in Europe have been more marked for large-bodied and wing less spec ies , as well as habitat and feeding special is ts (Kotze and O 'Hara 2003). However, when provided with habitat nearby, recolonizat ion of previously harvested forest by larger spec ies of carabids may happen rapidly (Brouat et al. 2004). In order to understand how carabid populat ions are affected by forest harvest ing, a more detai led understanding of how populat ion dynamics change with disturbance is required. In addit ion to studying the number of an imals found in different habitats, movement of a spec ies across a landscape adds important information to our understanding of population p rocesses such as migration and perception of habitat (Schoo ley and W iens 2003). Beyond tracing movements in order to plot population changes , movement dynamics are a useful tool to interpret use and infer habitat suitability. For example , the d is tances covered by individuals have been found to vary greatly with habitat suitability, with individuals travelling further and faster in inhospitable habitat (Baars 1979). A l so , the number of turns an animal makes in its chosen trajectory (i.e. path tortuosity) has been correlated in a number of spec ies with different behaviours such as foraging and migratory movement (R ieken and Raths 1996, Charr ier et al. 1997). A s with populat ion dynamics studies, behavioural studies on forest-dwell ing insects are not common in the forest management setting, inhibiting their inclusion into forest management planning (Loye and F isher 1999). 5 In addit ion to studying the abundance of spec ies and the movement within populat ions, a c loser examinat ion of the food resources avai lable in a variety of habitats is necessary . Before invoking mechan i sms such a s food limitation, a better understanding of the mechan isms for abundance and distribution of potential food spec ies is important (Jul iano and Lawton 1990). E v e n if food is not a limiting factor to a spec ies , the distribution of food resources may determine where it is found (Van de Koppe l et al. 2005) In order to advance our understanding of the mechan isms affecting populat ions, we must first examine the effects on individuals within the population. Physio log ica l responses to environmental condit ions may manifest themselves as changes in fecundity (Perrin and S ib ley 1993), migration rates (Dixon etal. 1993), or death rates, but they will first become apparent in the condit ion of individuals (Perrin 1991). In vertebrates, immediate responses of the individual to stress can be measured via the re lease of s o m e steroid hormones (Ricklefs and Wike lsk i 2002). This and other uses of physiological responses to environmental s t ressors are gaining popularity in quest ions regarding the conservat ion of spec ies , termed "conservat ion physiology" by Wikelsk i and C o o k e (2006). Whi le such endocr ine sys tems are wel l -understood in vertebrates, no ana logues are known for insects (Lovejoy and J a h a n 2006). A s proxy measures to this, internal condit ions are good measures of past physiological condit ions (Ricklefs and Wike lsk i 2002). Measu res of internal condit ions such as energy storage, gut contents, reproductive development and production of young (eggs) are all important upon which to make inference about the mechan isms that drive populat ions (Perrin 1991). Internal condit ions may be used to evaluate habitat suitability for a spec ies 6 (Ostman 2005), and may indicate the p resence of food resources that are specif ic to undisturbed, older habitats (Barone and Frank 2003). Suitability of a spec ies to a particular habitat may be reflected in internal condit ions, but caution should be used when drawing tighter assoc ia t ions between c a u s e s and effects. Un less individuals are known to have inhabited a particular habitat type for a relevant amount of t ime, their assoc iat ion with that habitat may be owing to migration and their internal condit ion may not be indicative of the habitat's suitability for them. Similarly, individuals with certain internal condit ions (e.g. very low feeding levels) may be more likely to migrate from an unsuitable habitat (Baars 1979), resulting in an overest imat ion of a habitat's suitability for a spec ies . Ques t ions With much of riparian forest management centred on vertebrates and aquatic organ isms, the impacts on other terrestrial spec ies need to be quantif ied and the eco logy of spec ies in this habitat better understood. Do riparian reserves retain similar spec ies to riparian a reas of cont iguous forest? If similar spec ies are found in these areas , do they perceive their environment in the same manner? Do their interactions with their environment c h a n g e ? What mechan isms might be responsib le for any changes in the ecology of terrestrial spec ies in these habitats? A s indicators of habitat d isturbance, what aspects of the ecology of carabid beet les are affected by harvest ing in riparian a reas? How useful are labour-intensive methods of evaluat ing internal condit ions a s compared to external metr ics? 7 In this thesis, I explore these quest ions as they pertain to the carabid beetle, Scaphinotus angusticollis, and evaluate this fl ightless, forest-dwell ing insect as a focal spec ies for further studies. S tudy Organ ism I se lected the carabid beetle, Scaphinotus angusticollis (F ischer V o n Waldhe im) a s a focal spec ies for this study for severa l reasons. A s a fl ightless and large beetle, S . angusticollis w a s conduc ive to mark-recapture study as well as for examinat ion of internal condit ions such as gut ful lness and fat body s ize . Scaphinotus angusticollis also compr ises a large proportion of the terrestrial insect b iomass and accounts for about 7 0 % of the individual carabid beet les caught in some coasta l forests of British Co lumb ia (R ichardson et al. in prep.). Scaphinotus angusticollis prefers to live in damp, forested habitats of western North Amer i ca (Lindroth 1961) and studies with before versus after compar isons have shown that populat ions decl ine in forested a reas that were harvested (Lenski 1982, Lemieux and Lindgren 2004). Scaphinotus angusticollis is thought to spec ia l ize in eating snai ls (Lindroth 1961, Thie le 1977). Little is known about the life history of S . angusticollis, but the larvae have been identified (Thiele 1977). Th is research w a s conducted at the Malco lm Knapp Resea rch Forest in Maple R idge, B C . Chapter Two of this thesis evaluates the relative precis ion of three methods for estimating population s izes of terrestrial arthropods. The third chapter compares population s i zes and movement behaviour of Scaphinotus angusticollis in three different riparian habitat types. In order to quantify potential food resources used by S . angusticollis in the s a m e riparian habitats, Chapter Four of this thesis examines the 8 abundance and diversity of snai ls , thought to be their main prey item. Chapter Five examines the internal condit ions of S . angusticollis in the three riparian habitat types, reveal ing potential mechan isms that drive this beetle's populat ions. Literature Ci ted Anonymous . 1995. Ripar ian Management A r e a Guidebook. British Co lumb ia Ministry of Environment and British Co lumb ia Ministry of Forests . Forest Pract ices C o d e of Brit ish Co lumb ia . Baars , M.A. 1979. Patterns of Movement in Radioact ive Carab id Beet les. Oeco log ia 44: 125-140. Barone, M. and T. Frank. 2003. Habitat age increases reproduction and nutritional condit ion in a general ist arthropod predator. Oeco log ia 135; 78-83. Boul ton, A . M . and K.D. Amberman . 2006. How ant nests increase soi l biota r ichness and abundance: a field experiment. Biodiversity and Conservat ion 15: 69-82. Brouat, C , S . Meusnier , and J . -Y . Rasp lus . 2004. Impact of forest management pract ices on carabids in European fir forests. Forestry 22: 85-97. Ca ro , T . M . and G . O'Doherty. 1999. O n the use of surrogate spec ies in conservat ion biology. Conservat ion Biology 13: 805-814. Charr ier, S . , S . Petit, and F. Burel . 1997. Movements of Abax parallelepipedus (Coleoptera, Carab idae) in woody habitats of a hedgerow network landscape: a radio-tracing study. Agriculture, Ecosys tems , and Environment 61 : 133-144. Col l inge, S . K . , M. Holyoak, C . B . Barr, and J .T. Marty. 2001 . Ripar ian habitat fragmentation and population pers istence of the threatened val ley elderberry longhorn beetle in central Cal i fornia. Biological Conservat ion 100: 103-113. Cummins , K .W. and M.A. W i l zbach . 2005 . The inadequacy of the f ish-bearing criterion for. st ream management . Aquat ic S c i e n c e 67: 486-491. Dixon, A . F . G . , S . Horth, and P. K ind lemann. 1993. Migration in insects: costs and strategies. Journal of An ima l Eco logy 62: 182-190. Dunn, R .R . 2005. Modern insect extinctions, the neglected majority. Conservat ion Biology 19: 1030-1036. G r e z , A . , T. Zav iezo , L. Tischendorf , and L. Fahr ig. 2004. A transient, positive effect of habitat fragmentation on insect population densit ies. Oeco log ia 141: 444-451. Grove , S . J . 2002. Saproxyl ic insect eco logy and the sustainable management of forests. Annua l Rev iew of Eco logy and Systemat ics 33: 1-23. Grove , S . and B. Yax ley . 2005. Wildlife habitat strips and native forest ground-act ive beetle assemb lages in plantation nodes in northeast T a s m a n i a . Harris, G . M . and S .L . P i m m . 2004. Bird spec ies ' tolerance of secondary forest habitats and its effects on extinction. Conservat ion Biology 18: 1607-1616. Her ing, D. 1998. Ripar ian beet les (Coleoptera) along a smal l s t ream in the Oregon Coas t Range and their interactions with the aquatic environment. The Coleopter ists Bulletin 52: 161-170. Hodk inson, I.D. and J .K . J a c k s o n . 2005. Terrestrial and aquatic invertebrates as bioindicators for environmental monitoring, with particular reference to mountain ecosys tems : Environmental Management 35: 649-666. Hylander, K., C . N i lsson, and T. Gothner. 2004. Effects of buffer-strip retention and clearcutt ing on land snai ls in boreal riparian forests. Conservat ion Biology 18: 1052-1062. 10 Hyvar inen, E. , J . Kouk i , P. Mart ikainen, H. Lappa la inen. 2005. Short-term effects of controlled burning and green-tree retention on beetle (Coleoptera) assemb lages in managed boreal forests. Forest Eco logy and Management 212: 315-332. J o n s s o n , M., T. Ran ius , H. Ekval l , G . Bostedt, A . Dahlberg, B. Ehnst rom, B. Norden, and J . N . Stok land. 2006. Cost-ef fect iveness of silvicultural measures to increase substrate availabil ity for red-listed wood-l iving organ isms in Norway spruce forests. Biological Conservat ion 127: 443-462. Ju l iano, S . A . and J . H . Lawton. 1990. Extr insic vs. intrinsic food shortage and the strength of feeding links: effects of density and food availability on feeding rate of Hyphydrus ovatus. Oeco log ia 83 : 535-540. K l imaszewsk i , J . , D.W. Langor, T.T. Work, G . Pelletier, H . E . J . Hammond , and C . Germa in . 2005. The effects of patch harvesting and site preparation on ground beetles (Coleoptera, Carab idae) in yel low birch dominated forests of southeastern Quebec . Canad ian Journal of Forest Resea rch 35: 2616-2628. Koivula, M. and J . N iemela . 2003 . G a p felling as a forest harvesting method in boreal forests: responses of carabid beet les (Coleoptera, Carab idae) . Ecography 26: 179-187. Kotze , D .J . and R .B . O 'Hara . 2003 . S p e c i e s decl ine - but why? Explanat ions of carabid beetle (Coleoptera, Carab idae) dec l ines in Europe. Oeco log ia 135: 138-148. L e e , P. and T. Barker , 2005. Impact of riparian buffer guidel ines on old growth in western boreal forests of C a n a d a . Forestry 78: 263-278. Lemieux, J . P . and B . S . Lindgren. 2004. Ground beetle responses to patch retention harvesting in high elevation forests of British Co lumb ia . Ecography 27: 557-565. 11 Lensk i , R . E . 1982. The impact of forest cutting on the diversity of ground beet les (Coleoptera: Carab idae) in the southern Appa lach ians . Ecolog ica l Entomology 7: 385-390. Lindroth, C . H . 1961. The Ground Beet les of C a n a d a and A l a s k a . Opuscu la Entomologica Supp lementum X X - X X I V . P p . 1-1192. Lovejoy, D.A. and S . J a h a n . 2006. Phytogeny of the cort icotrophin-releasing factor family of peptides in the metazoa . Genera l and Comparat ive Endocr inology 146 1-8. Loye, J . and B. Fisher. 1999. Importance of insect behavior in conservat ion. Anna ls of the Entomological Soc iety of Amer i ca 92: iv. Moore , J . W . 2006. An imal ecosys tem engineers in streams. B i o s c i e n c e 56: 237-246. Moore, R.D. 2005. Smal l s t ream channe ls and their riparian zones in forested catchments of the Paci f ic Northwest: Introduction. Journal of the Amer i can Water Resou rces Assoc ia t ion 41 : 759-761. Moore, R .D. and J . S . R ichardson . 2003. Progress towards understanding the structure function, and ecological s igni f icance of smal l stream channe ls and their riparian zones . Canad ian Journal of Forest Resea rch 33 : 1349-1351. Nakano, S . , H. M iyasaka , and N. Kuhara . 1999. Terrestr ial-aquatic l inkages: Ripar ian arthropod inputs alter trophic c a s c a d e s in a stream food web. Eco logy 80: 2435-2441 . Oliver, I., R. M a c Nally, and A . York. 2000. Identifying performance indicators of the effects of forest management on ground-act ive arthropod biodiversity using 12 hierarchical partitioning and partial canonica l cor respondence analys is . Forest Eco logy and Management 139: 21-40. O s a w a , N., A . Tera i , K. Hirata, A . Nakanish i , A . Makino, S . S a k a i , and S . S iba ta . 2005. Logging impacts on forest carabid assemb lages in J a p a n . Canad ian Journa l of Forest Resea rch 35: 2698-2708. Os tman , O. 2005. Asynchronous temporal variation among sites in condit ion of two carabid spec ies . Ecolog ica l Entomology 30: 63-69. Pae tzo ld , A . and K. Tockner . 2005. Effects of riparian arthropod predation on the b iomass and abundance of aquatic insect emergence. Journal of the North Amer ican Benthological Soc iety 24: 395-402. Perr in, N. 1991. Opt imal resource al location and the marginal value of organs. T h e Amer ican Naturalist 139: 1344-1369. Perr in, N. and R . M . Sib ley. 1993. Dynamic models of energy al location and investment. Annua l Rev iew of Eco logy and Systemat ics 24: 379-410. Po lasky , S . , E. Ne lson , E. Lonsdorf, P. Fackler , and A . Starf ield. 2005. Conserv ing spec ies in a working landscape: land use with biological and economic object ives. Ecolog ica l Appl icat ions 15: 1387-1401. Potvin, F., N. Bertrand, and J . Ferron. 2005. Attributes of forest strips used by snowshoe hare in winter within clear-cut boreal landscapes . Canad ian Journa l of Forestry Resea rch 35: 2521-2527. Rain io , J . and J . N iemela . 2002. Ground beet les (Coleoptera: Carab idae) as bioindicators. Biodiversity and Conservat ion 12: 487-506. 13 Richardson, J . S . , R . J . Na iman, F . J . S w a n s o n and D.E. Hibbs. 2005. Ripar ian communit ies assoc ia ted with Paci f ic Northwest headwater s t reams: assemb lages , p rocesses , and uniqueness. Journa l of the Amer i can Wate r Resou rces Assoc ia t ion 41 : 935-947. Ricklefs, R . E . and M. Wikelsk i . 2002. The physiology/l i fe history nexus. T rends in Eco logy and Evolut ion 17: 462-468. R iecken , U. and U. Raths. 1996. U s e of radio telemetry for studying d ispersa l and habitat use of Carabus coriaceus L. Annates Zoologic i Fennic i 33 : 109-116. Roderwald , A . D . and M.H. Bakermans . 2006. What is the appropriate paradigm for riparian forest conservat ion? Biological Conservat ion 128: 193-200. S a b o , J .L . , R. Sponsel ler , M. Dixon, K. G a d e , T. Harms, J . Herrernan, A . Jan i , G . Katz , C . Soykan , J . Watts, and J . Welter. 2005. Ripar ian z o n e s increase regional spec ies r ichness by harboring different, not more, spec ies . Eco logy 86: 56-62. Schoo ley , R.L. and J .A . Wiens . 2003. Finding habitat patches and directional connectivity. O ikos 102: 559-570. Shir ley, S . M . 2006. Movement of forest birds across river and clearcut edges of varying riparian buffer strip widths. Forest Eco logy and Management 223 : 190-199. Simberloff, D. 1999. The role of sc ience in the preservat ion of forest biodiversity. Forest Eco logy and Management 115: 101-111. S w a n s o n , F . J . and J . F . Frankl in. 1991. New forestry principles from ecosys tem analys is of Paci f ic Northwest forests. Eco log ica l Appl icat ions 2: 262-274. 14 Symondson , W . O . C . , K.D. Sunder land, and M .H . Greens tone . 2002 . C a n general ist predators be effective biocontrol agents? Annua l Rev iew of Entomology 47: 561 -94. Taylor, R . J . and N. Doran. 2001 . U s e of terrestrial invertebrates a s indicators of the ecological sustainabil i ty of forest management under the Montreal P r o c e s s . Journal of Insect Conservat ion 5: 221-231 . Thie le, H .U. 1977. Carab id Beet les in their Envi ronments. Spr inger Ver lag , Berl in. Pp . xvi i -369. V a n Looy, K., S . Vanacker , H. J o c h e m s , G . DeBlust , and M. Dufrene. 2005. Ground beetle habitat templets and riverbank integrity. River R e s e a r c h and Appl icat ions 21 :1133 -1146 . V a n de Koppe l , J . , R.D. Bardgett, J . Beng tsson , C . Rodr iguez-Bar rueco , M. Rietkerk, M .J . W a s s e n , and V . Wol ters. 2005. The effects of spatial sca le on trophic interactions. Ecosys tems 8: 801-807. Wikelsk i , M. and S . J . C o o k e . 2006. Conservat ion physiology. T rends in Eco logy and Evolut ion 21 : 38-46. Wipfl i , M. 2005. Trophic l inkages between headwater forests and downst ream fish habitats: implications for forest and fish management . L a n d s c a p e and Urban Planning 72: 205-213. Work, T.T, J . R . S p e n c e , W . J . A . Volney, L .E . Morgant ini , and J .L . Innes. 2003. Integrating biodiversity and forestry pract ices in western C a n a d a . The Forestry Chronic le 79: 906-916. 15 Chapter Two Evaluation of two population estimate methods for ground beetle (O. Coleoptera, F. Carabidae) populations Introduction Populat ion s ize est imates are der ived using many different methods, with some employing more information (e.g. recapture rate, movement, age structure, and effort of sampl ing) than others (e.g. p resence /absence recordings). For ecologists trying to a s s e s s the s izes as well as spatial and temporal abundance patterns of wild populat ions, accuracy and the amount of time required are often the key factors used to select a protocol for sampl ing or monitoring (Casagrande and Be iss inger 1997; Fo rcada 2000; Bl igaard 2001). Depending on the type of information that is required, est imates of population s izes can be obtained by using different models . Extensive models that use detai led information are cons idered to give the most accurate assessmen t of population s ize, but are frequently avo ided in favour of less demand ing methods (Melbourne 1999). In a recent review of the var ious models of estimating populat ions, S c h w a r z and Sebe r (1999) indicate that there needs to be s o m e study of the efficiency in deriving population est imates from different models . Catch-per-unit-effort ( C P U E ) est imates of population s ize are the most commonly used form of population est imate in insect ecology, appear ing in many forms such as indices of transect counts and relative abundance est imates. T h e s e are the lowest-effort form of population est imates and provide the least amount of information about the individuals, merely recording an individual 's p resence and level of activity in a particular area. Est imates of population s ize can also be based on capture-recapture data. There are many different mode ls for est imating populat ion s ize from capture-16 recapture data; a recent review by S c h w a r z and S e b e r (1999) summar ized current deve lopments in this f ield, including the use of logistic regression, l ikelihood models and Bayes ian analys is to improve populat ion est imates from commonly -used wildlife monitoring methods. The use of different types of populat ion est imates is facilitated by a number of factors that are determined by the biology and behaviour of the organism. Larger animals that are active during daylight are eas ier to monitor using capture-recapture methods. More abundant, nocturnal an imals are more easi ly monitored using catch-kil l sampl ing. The number of an imals caught in this kind of trapping is converted to a catch-per-unit-effort ( C P U E ) est imate of population s ize. Typical ly, terrestrial arthropod studies have only employed C P U E est imates of population s ize, although there are s o m e notable except ions in mark-recapture studies on darkl ing beet les and butterflies (Mclntyre and W iens 1999, Haddad 2000, Ro land et al. 2000). Owing to low recapture rates and effort invested in marking individuals, capture-recapture population est imates are not typically used . Different population est imates are rarely calculated from the s a m e population at the same t ime. If the different populat ion est imates are poor approximations of one another, they should generate different est imates for the s a m e populat ions at the same time. G iven that catch-per-unit-effort est imations are based on a proportion of the populat ion, I hypothesize that they underest imate population s izes as compared to mark-recapture methods. Further, as populat ions in different habitats were evaluated, prior studies predict that different populat ion est imates should vary more from one 17 another, a s est imates are more likely to be affected by movement, vegetat ion and microcl imate. The value in compar ing severa l s imul taneous est imates highlights the relative precis ion of the population est imates without the confounding factors that using different sites might pose. The objective of this study w a s to evaluate how well commonly -used measures of relative abundance , like C P U E , approximate the more labour-intensive populat ion est imates such as those derived from mark-recapture methods. W e compared two methods of population est imates for the same populat ions of ground beet les in an open setting. Resea rche rs frequently make the assumpt ion that while there is a sacri f ice in precis ion, catch-per-unit-effort population est imates are the only method feasib le for most arthropod studies (Melbourne 1999). I examined dif ferences between population est imates from mark-recapture and C P U E est imates to determine precis ion. Methods The nine study si tes for this experiment were establ ished along smal l headwater (first order) s t reams in the University of British Co lumbia 's Ma lco lm Knapp Resea rch Forest in Map le R idge, B C . The wet, coasta l forest at the study sites was logged in 1931 and suffered a w idesca le burn in the s a m e year and has s ince naturally regenerated. More recent logging took place from 1997 to 1999 as a part of a study on the effects of forest harvesting in riparian areas (Kiffney et al. 2003). For my examinat ion of a carabid beetle population, two levels of harvesting intensity in riparian a reas were examined, with three repl icates for each type of d isturbance: completely logged to the edge of the stream (no buffers or clearcut) and with 30 m of forest left on 18 either s ide of the st ream (30 m buffer strips). Three more sites provided exper imental controls with no logging. Pitfall traps were laid out to sample an area 45 m x 45 m immediately adjacent to the s t reams in each site in evenly spaced rows and co lumns, with each trap 5 m apart, i.e., 10 rows x 10 co lumns = 100 traps. T h e s e pitfall traps consis ted of two plastic cups embedded in the ground, one inside the other. To prevent rain from filling the cups , a 30 c m x 30 cm piece of plywood was laid over top, resting on four 15 c m long p ieces of 3 c m x 3 cm wood laid on the ground to act as guides for walking insects (Morrill et al. 1990). This spac ing w a s determined as the approximate minimum distance of travel for an individual beetle in one 12-hour period (R iecken and Raths 1996, Charr ier e r a / . 1997). These trapping grids were also des igned to be larger than the max imum dis tance travelled in 3 days by spec ies that prefer to dwell in wood lands (R iecken and Ra ths 1996, Charr ier et al. 1997). I carried out live trapping of Scaphinotus angusticollis (F ischer V o n Waldhe im) using pitfall traps that lacked any killing agent. New individuals trapped overnight in pitfall trap arrays were marked with a unique, three-digit number hand-painted on the elytra and re leased after weighing and measurement of their pronotum width. Identities of recaptured individuals were recorded, along with location, and body measurements were retaken to test for bias in researchers. Sequent ia l daily trapping col lect ions were conducted over a four-day period, giving three nights of trapping in a sess ion . Two different types of population est imates were generated for analys is: mean number caught per trapping week and capture-recapture est imates. M e a n number 19 caught per trapping week was back-calculated from capture-recapture data and is the equivalent of catch-per-unit-effort est imates of population s ize . B e c a u s e the length of each trapping period w a s short relative to the l i fespan of individuals and their est imated d ispersa l capabil i t ies, populat ions were cons idered using a "c losed" (no migration, births or deaths during the sampl ing) population model (Pol lock e r a / . 1990). Capture-recapture est imates were done using the S c h n a b e l -S c h u m a c h e r estimator (Krebs 1999) because of its ability to est imate a c losed populat ion's s ize (no migration, deaths or births) with the added information of repeated sampl ing. In addit ion to these two population est imators, a corrected S c h n a b e l -S c h u m a c h e r est imate w a s calculated, using mean amount of movement in e a c h habitat to correct for the effective grid s ize (Van H o m e 1982). The mean amount of movement w a s calculated by examining recapture data for each month and taking the mean distance travelled by recaptures in that time period. It w a s necessary to pool data for each habitat by month, as some habitat types had very few recaptures in some months. Trapping w a s conducted from mid-July 2000 to mid-September 2000 and again from mid-May 2001 to late August 2001 . In 2000, five sites were trapped in one week and four the next week, al lowing for a two-week rotation through all the sites. In 2001, three sites were trapped during most sess ions , giving a three-week rotation through all si tes. O n e replicate of each habitat type was sampled in each trapping sess ion , to minimize variation in results owing to seasona l or other temporal effects (e.g. weather patterns). During some time periods, there were no recaptures made and therefore no Schnabe l -Schumache r est imates for population s ize and no compar isons could be made. C P U E est imates were calculated from Exce l spreadsheets of the raw data, while 20 Schnabe l -Schumache r est imates were calculated using the statistical analys is package in Eco log ica l Methodology (Krebs, 1999). Data analys is Compar i sons were made between the different methods of density est imates using a concordance correlation analys is deve loped by Lin (1989, 2000). Th is method w a s chosen over other correlation coefficients (e.g. Pearson ' s , Kendal l 's) and least squares analys is , because it is specif ical ly des igned to analyze compar isons between two measures of the s a m e factor (Lin 1989). Th is analys is is enhanced by the assumpt ion that two measures of the same population or sample should be identical (Lin 1989). The range examined for analys is is similar for all compar isons made here, satisfying the requirement for the use of a similar analytical range made by Lin and Chinchi l l i (1997). Calculat ion of the concordance correlation analys is was done through the National Institute of Water and Atmospher ic R e s e a r c h (NIWA) websi te (www.niwa.co.nz/services/stat ist ical /concordance), which generates both a concordance correlation coefficient and conf idence intervals for this est imate. For the three logging types examined, concordance correlation coefficients were calculated for compar isons between population est imates of C P U E and each of the two types of Schnabe l -Schumache r est imates. Pool ing data from different harvesting patterns al lowed for the assessment of forest harvesting effects on population est imates. Pool ing data within each year al lows for the assessmen t of the effects of longer-term natural phenomena like weather patterns on population est imates. Statistical tests of di f ferences between concordance correlation va lues were made using a modified X 2 21 procedure (Zar 1996), and pairwise compar isons between individual concordance correlation values were made. Resul ts Not surprisingly, catch-per-unit-effort ( C P U E ) population est imates and Schnabe l -Schumache r est imates follow a l inear trend (Figure 2.1a), as do C P U E and corrected Schnabe l -Schumache r population est imates (Figure 2.1b). In both f igures, the d ispers ion of the data grows as population s ize est imates increase. Compar i son of F igures 2.1a and 2.1b shows that the d ispers ion of the data is greater in the corrected Schnabe l -Schumache r and C P U E figure and that the d ispers ion occurs over a greater portion of the range of est imates. These results initially suggest that even at the lower population densit ies observed in this study, C P U E population est imates are poor approximations of corrected Schnabe l -Schumache r est imates. A compar ison between Schnabe l -Schumache r and corrected Schnabe l -Schumache r indices (Figure 2.1c) il lustrates the effect of incorporating movement into population est imates. The data fall below the 1:1 trendline, indicating that Schnabe l -Schumache r est imates are consistently higher than corrected Schnabe l -Schumache r est imates. F igures 2.1a through 2.1c indicate the general spread of the est imates and where the greatest d iscrepanc ies lie with respect to population s ize , but the concordance correlation analys is al lowed for an objective analys is of the data spread and the change of any l inear relat ionships from a 1:1 ratio. In the compar ison of C P U E and Schnabe l -Schumache r est imates, control si tes had the lowest concordance value (Figure 2.2), indicating the greatest amount of d ispers ion between these two population est imates. At the other extreme of the 22 habitats examined, clearcuts had the highest concordance value (Figure 2.2) and therefore had the best fit of these population est imates to one another. However, there were no statistically significant di f ferences between habitats (X 2 caic= 0.340, X V 0.05)= ( 5.991). For the compar ison of C P U E and corrected Schnabe l -Shumache r est imates, a different pattern was evident in the concordance va lues. Whi le control and clearcut si tes had the lowest concordance va lues and therefore the greatest amount of d ispers ion, 30 m buffer sites had the greatest concordance and therefore the least amount of d ispers ion in population est imates. For the 30 m buffer habitat, concordance conf idence intervals over lapped with the other two concordance est imates. The X 2 analys is conf i rmed that there were no significant dif ferences in heterogeneity between habitats (X 2 ca ic= 1-134, ^ 2 ( 2 , 0 . 0 5 ) = 5.991). It is interesting to note that the concordance between population est imates was greatly increased with the inclusion of movement into the Schnabe l -Schumache r est imates (Figures 2.2 and 2.3), although the d ispers ion of the data did appear to be higher when visual ly compar ing Figures 2.1a and 2.1b. Th is is because the concordance correlation analysts evaluates not only the dispersion of the data from the l inear trend, but a lso the deviation of the linear trend from that of a 1:1 ratio line (Lin 1989) (as indicated in Figures 2.1a and 2.1b). To test for the effect of year on population est imates, est imates were pooled for e a c h year (Figure 2.3) and then compared using the X2 analys is. C P U E and corrected Schnabe l -Schumache r est imates did consistently show the highest amount of concordance and C P U E and Schnabe l -Schumache r est imates showed the lowest 23 amount of concordance (Figure 2.3). However, according to the X 2 results, there w a s no significant difference in concordance between years (X 2 caic= 2.416, X V o.o5)= 11.070). In order to further investigate where the largest di f ferences occurred, pairwise Tukey test compar isons by habitat were performed according to Za r (1996) and recommended by Lin (1989). Of the three habitats examined in this study, the 30 m buffer habitats showed the greatest difference in concordance results (qcaic=1 -391, qCrit=2.772), but this is still not a statistically significant difference. D iscuss ion Accura te , inexpensive and practical methods of assess ing population numbers are crucial to ecologists. Frequently in insect eco logy studies, obtaining catch-per-unit-effort population est imates is preferred to more intensive means like capture-recapture methods, although there are a number of notable except ions. In this study, I test if it is necessary to invest the extra effort required for capture-recapture population est imates for terrestrial invertebrates. Or: A re capture-recapture population est imates substantial ly different from standard catch-per-unit-effort es t imates? The results of this study indicate that while there is certainly a high degree of variability in population est imates, C P U E and capture-recapture est imates follow a similar, c lear and predictable trend. Whi le the fit of these est imates to one another would only be considered "fair" by biomedical s tandards (0.21 to 0.40, N IWA websi te), the concordance between population est imates was remarkable, g iven the high degree of variability inherent to field ecology. Conco rdance in the case of the compar ison between C P U E and corrected Schnabe l -Schumache r est imates in the 30 m buffer si tes 24 (0.648) was in the range considered to be "substantial" in b iomedical research (NIWA website), indicating a remarkably good fit of these two est imators to one another. Compar i son of the population est imates a lso indicates that there is a strong trend fol lowed by est imates over a range of population s izes observed . Conco rdance analys is showed some unexpected results of using the different population est imators. Whi le the initial prediction w a s that the adjustment of S c h n a b e l - S c h u m a c h e r population est imates using movement data would cause further data d ispers ion, it actual ly improved the fit of the population est imates to that of the C P U E est imates, as ev idenced in the higher concordance va lues obtained in the compar ison of C P U E with corrected Schnabe l -Schumache r est imates. Th is means that the most detai led model of estimating population s ize (corrected Schnabe l -Schumacher ) is the c losest fit to C P U E est imates. In more practical terms, population est imates made using C P U E require the least amount of effort and are a good approximation of the more detai led model . I tested for the potential effects of yearly di f ferences in environmental factors on population est imates. Al though any dif ferences between the two years could be attributed to a range of factors, from weather pattern variation to sampl ing effects, no significant dif ferences were observed between the two years , as determined by the X2 analys is . This means that although prior studies have identified cl imate as a problematic source of variation in catchabil i ty (Raworth and Cho i 2001), for the purposes of this study, no addit ional heterogeneity in populat ion est imates w a s noted. Whi le this does not suggest that microcl imate does not play a role in determining local distributions, it does suggest that yearly variation in weather patterns (in addit ion to the 25 effects of population sampl ing over prolonged per iods of t ime) does not ser iously bias different methods of population est imates of terrestrial arthropods, Perhaps the most interesting finding of this study lies within the compar isons made for individual harvesting types. There w a s less d ispers ion (i.e. higher concordance values) in the population est imators in clearcut habitats than in control (uncut) habitats. After correction for movement d ispers ion w a s lessened (i.e. concordance results almost doubled in value) in populat ion est imates for control and 30 m buffer habitats (Figure 2.2), but remain relatively similar in clearcut habitats. Whi le statistical compar ison of concordance va lues did not show any signif icant di f ferences, results for the 30 m buffer si tes were nearly signif icant. Th is indicates that habitat may have some effect on the precision of population est imates. In the 30 m buffer si tes, population est imates are altered more by the inclusion of movement within the population than the other two habitats. Th is is of concern to ecologists attempting to estimate population s izes , as compar isons between habitats are often integral to quest ions posed about a spec ies ' response. Al though there were di f ferences in the dispersion of the population est imates because movement w a s included, the overal l coefficient of concordance correlation was not significantly different between the three habitats observed. T h e s e results parallel some work on other forest-dwell ing taxa, indicating that corridors of forested habitat may be used for d ispersal movement (Hannon etal. 2002, Haddad et al. 2003 , Monkkonen and Mutanen 2003 , Varkonyi et al. 2003). Catch-per-unit-effort est imates ( C P U E ) were good predictors of the more detai led population est imates, regardless of habitat and year. Th is g ives ev idence that, for 26 assessmen t of wild, open populat ions of terrestrial arthropods, C P U E est imates are best investment when the full considerat ion of effort and est imate precision is made. 27 Literature Ci ted Bl igaard, J . 2001 . Binomial sampl ing as a cost efficient sampl ing method for pest management of cabbage root fly (Dipt., Anthomyi idae) in caulif lower. Journa l of App l ied Entomology 125: 155-159. Casag rande , D . G . and S . R . Beiss inger . 1997. Evaluat ion of four methods for estimating parrot population s ize . The Condo r 99: 445-457. Charr ier, S . , S . Petit, and F. Bure l . 1997. Movements of Abax parallelepipedus (Coleoptera, Carab idae) in woody habitats of a hedgerow network landscape: a radio-tracing study. Agriculture, Ecosys tems , and Environment 61 : 133-144. Fo rcada , J . 2000. C a n population surveys show if the Medi terranean Monk S e a l colony at C a p B lanc is decl ining in abundance? Journal of Appl ied Eco logy 37: 171-181. Haddad , N. 2000. Corr idor length and patch colonizat ion by a butterfly, Junonia coenia. Conservat ion Biology 14: 738-745. Haddad , N .M. , D.R. Bowne, A . Cunn ingham, B . J . Dan ie lson, D.J . Levey, S . Sargent , and T. Sp i ra . 2003 . Corr idor use by diverse taxa. Eco logy 84: 609-615. Hannon , S . J . , C A . Paszkowsk i , S . Bout in, J . DeGroot , S . E . Macdona ld , M. Wheat ley , and B.R. Eaton . 2002. Abundance and spec ies composi t ion of amphib ians, smal l mammals , and songbirds in riparian forest buffer strips of varying widths in the boreal mixedwood of Alber ta. C a n a d i a n Journal of Forest Resea rch 32:1784-1800. Kiffney, P . M . , J . S . R ichardson, and J . P . Bul l . 2003. R e s p o n s e s of periphyton and insects to manipulat ion of riparian buffer width a long forest s t reams. Journal of Appl ied Eco logy 40: 1060-1076. 28 Krebs , C . J . 1999. Eco log ica l Methodology. Ben jamin /Cummings : Don Mil ls, O N . Lin, L. I.-K. 1989. A Conco rdance Correlat ion Coeff icient to Evaluate Reproducibi l i ty. Biometr ics 45 : 255-266. L in, L. I.-K. 2000. A note on the Conco rdance Correlat ion Coefficient. Biometr ics 56: 324-325. L in, L. I.-K. and V . Chinchi l l i . 1997. A Rejoinder to the Letter to the Editor from Atk inson and Nevi l l . Biometr ics 53: 777-778. Mclntyre, N .E . and J .A . W iens . 1999. How does habitat patch s ize affect animal movement? A n experiment with darkl ing beet les. Eco logy 80: 2261-2270. M a c K e n z i e , D.I. and W . L . Kendal l . 2002. How should detection probability be incorporated into est imates of relative abundance? Eco logy 83 : 2387-2393. Melbourne, B.A. 1999. B ias in the effect of habitat structure on pitfall traps: A n exper imental evaluat ion. Austral ian Journal of Zoology 24: 228-239. Monkkonen , M. and M. Mutanen. 2003 . Occu rence of Moths in Borea l Forest Corr idors. Conservat ion Biology 17: 468-475. Morril l, W .L . , D . G . Lester, and A . E . Wrona . 1990. Factors affecting eff icacy of pitfall traps for beet les (Coleoptera, Carab idae and Tenebr ionidae). Journal of Entomological S c i e n c e 25 : 284-293. Pol lock, K .H . , J . D . Nichols, C . Brownie, and J . E . H ines. 1990. Statistical Inference for Capture-Recapture Exper iments. Wildlife Monographs 107: 1-97. Raworth, D.A. and M.-Y. C h o i . 2001 . Determining numbers of active carabid beet les per unit a rea from pitfall-trap data. Entomologia Experimental is et Appl icata 98: 95-108. 29 Riecken , U. and U. Raths . 1996. U s e of radio telemetry for studying d ispersal and habitat use of Carabus coriaceus L Anna les Zoologic i Fennic i 33 : 109-116. Ro land , J . , N. Keyghobad i , and S . Fownes . 2000. Alp ine Parnassius butterfly d ispersa l : Effects of landscape and population s ize . Eco logy 81 : 1642-1653. Schwarz , C . J . and G . A . F . Seber . 1999. Est imating animal abundance: Rev iew III. Statist ical S c i e n c e 14: 427-456. V a n H o m e , B. 1982. Effective trapped area for live-trap grids. Journal of Mammology 63:155-157. Varkony i , G . , M. Kuussaar i , and H. Lappa la inen. 2003. U s e of forest corridors by boreal Xestia moths. Oeco log ia 137:466-474. Zar , J . H . 1996. Biostatist ical Ana lys is . Prent ice Hal l , New Jersey . 30 Figures CD X I CD ocr o.3o e fe 0.25 E S ° - 2 0 <D o 0.15 CD E E o.io 0.05 0.00 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00 0.14 0.12 0.10 A 0 H • Cont ra 0 30 m V Clearcu t H n • H / • • • 0.00 0.02 0.04 0.06 0.08 0.10 C P U E (number / m z ) CD o CD E CD 1 E o I W c •a v CD o O B H • m m • V •• V • B • 0.00 0.02 0.04 0.06 0.08 C P U E (number / m 2 ) 0.10 E 3 CD . O i -CD 9> E 3 a; o CD o O 0.04 0.00 c 7 / / S • / B B /-. w A " • / • • 0.00 0.05 0.10 0.15 0.20 0.25 0.30 Schnabe l -Schumacher (number / m 2 ) Figure 2.1 Comparison of population estimates for the adult carabid Scaphinotus angusticollis sampled in 2000 and 2001 in control, 30 m buffer and clearcut habitats. Estimates were made from pitfall trapping and individual marking of beetles. A) Schnabel-Schumacher and C P U E estimates; B) Corrected Schnabel-Schumacher and C P U E estimates; and C) Corrected Schnabel-Schumacher and Schnabel-Schumacher estimates. Dashed line represents 1:1 ratio. 31 1.2 1.0 c 0 O i t <D O O o __ CD I L o o CD O c TO T3 i O o c o o 0.8 A 0.6 0.4 0.2 -\ • C P U E vs S S • C P U E vs c S S 0.0 4> Treatment Figure 2.2 Concordance correlation results, comparing catch-per-unit-effort (CPUE) and Schnabel-Shumacher (SS) population estimates (circles) and catch-per-unit-effort (CPUE) and corrected Schnabel-Schumacher (cSS) population estimates (squares) in each of the three habitats: control, 30 m buffer and clearcut sites. -.1.0 .10.8 o it CD O _0.6 g __ 0 i _ i _ o o CD O c CO 1 o o c o ^0.0 0.4 0.2 C P U E versus corr Schnabel-Schumacher C P U E versus Schnabel-Schumacher Year Figure 2.3 Concordance correlation analysis results for all Scaphinotus angusticollis population estimates, pooled by year 32 Chapter Three Relative abundance and movement of the carabid beetle Scaphinotus angusticollis in coniferous riparian forests of the Pacific Northwest Introduction The study of how habitat loss and fragmentation affect spec ies is a major focus for many population ecologists. Many anthropogenic d is turbances such as agriculture, forestry and urbanization alter habitat and result in the loss of spec ies . Conservat ion efforts must examine way s in which the impact of development is mitigated by speci f ic types of management strategies (Hilderbrand et al. 2005) and how non-target spec ies may benefit from these strategies (Elphick 2004). Forest harvesting has been demonstrated to reduce populat ions of many forest-preferring spec ies , that typically occur in cont iguous forest. At smal ler sca les one of the ways forest managers can reduce the impact of harvesting on forest-preferring spec ies is to leave treed reserves embedded within harvested areas. For example, in the Paci f ic Northwest in genera l , these may take the form of wildlife tree patches or riparian reserves (Swanson and Frankl in 1991). Theoretical ly, if populat ions of forest-preferring spec ies can persist in forest remnants, then they might be able to recolonize the regenerating forest when the appropriate condit ions develop in the surrounding landscape (Koivula et al. 2002). Of particular concern to conservat ion efforts are the needs of less mobile spec ies that must be planned for and met on a smal ler-scale (Hylander et al. 2004, Matlack and Monde 2004). L e s s vagile spec ies may have difficulty re-establ ishing in the core of harvested areas, as they are precluded by distance (Hylander et al. 2004). The design of forest 33 reserves must suit the biological requirements of the spec ies they are intended to protect and the planning of reserves must occur at sca les relevant to the organ isms. Forest reserves des igned to buffer s t reams from land use pract ices are one example of smal ler -scale management that may help forest-preferring spec ies . Frequently, forest remnants left around st reams in the riparian zone are referred to as riparian buffers, because they are intended to buffer s t reams from the microcl imatic and physical changes assoc ia ted with logging (Moore et al. 2005). Whi le the des ign of riparian buffers is primarily constrained by requirements for f ish protection (Young 2000, Lee et al. 2004) and shading for s t reams (Naiman and D e c a m p s 1997), addit ional benefits to these forest remnants include providing habitat for forest-preferring birds (Darveau et al. 1995, Hannon and Schmiege low 2002, Shir ley and Smi th 2005) and smal l mammals (Cock le and R ichardson 2003, Potvin et al. 2005), or for movement corridors (Monkkonen and Mutanen 2003). Spec ies that use riparian buffers may be ones that are typically only found c lose to st reams (riparian) or may be spec ies that only use the area facultatively (non-obligate spec ies) (R ichardson et al. 2005). The needs of forest-preferring invertebrates in forest remnants and riparian areas has been little studied (Hylander etal. 2004) and most studies have focused on the use of riparian buffers by flying insects in logged landscapes (Haddad etal. 2003 , Monkkonen and Mutanen 2003). Habitat requirements are very different for flying insects that require open habitat for flight paths and (in the case of Lepidoptera) the presence of particular f lowering plants for nectaring (Haddad et al. 2003). B e c a u s e of these rather speci f ic habitat requirements for flying insects, any extension of these f indings to other forest-preferring invertebrate spec ies is tenuous (Monkkonen and 34 Mutanen 2003). In fact, the study of non-flying, forest-preferring insects represents a large gap in our understanding of how the needs of forest-preferring insects use riparian buffers, because many forest insects dwell solely on the ground (Varkonyi et al. 2003). In ecology, the most evident measure of a population's welfare is to study the s ize and growth of the population, but there are severa l other aspects that can be examined that may shed light on the role a habitat plays. It is important to approach the study of a population from severa l aspects , because simple measures such as population s i zes may be mis leading, especial ly when data are col lected over a short period of t ime. For example, movement within a population may indicate migration or d ispersal capabi l i t ies and patterns. Movement can be a s s e s s e d in many ways , including the distance travelled by individuals and the shape of the path taken. Whi le direct inferences for a mechan ism are problematic with movement patterns, any significant di f ferences in movement patterns between areas studied suggest that the perception of the habitats differs for the study subject (Szyszko et al. 2004). Carab id beetles (F. Carab idae) present an excel lent study subject for a number of reasons. Carab ids have been shown to be good bioindicators of habitat d isturbance (Rainio and N iemela 2002) and have wel l-known assoc ia t ions with microcl imatic condit ions (Spence and Niemela 1994). Within the carabids, spec ies in the Tr ibe Cychr in i are found in abundance in many forested landscapes (Lemieux and Lindgren 2004), and appear to decl ine with logging (Lenski 1982). Of the severa l potential study spec ies in this group, I have chosen a large, fl ightless carabid beetle (Scaphinotus angusticollis) to investigate the effects of riparian buffers on terrestrial insect populat ions. 35 Scaphinotus angusticollis compr ises a large proportion of the terrestrial insect b iomass and accounts for about 7 0 % of the individual carabid beet les caught in pitfall trap samp les at Malco lm Knapp Resea rch Forest in Map le R idge, B C (R ichardson et al. in prep.). Scaphinotus angusticollis is assoc ia ted with wet, forested habitats of C a n a d a and A l a s k a (Lindroth 1961) and has also been found to be sensi t ive to logging. S tud ies with before versus after compar isons have shown that populat ions of S . angusticollis decl ine in forested areas that have been harvested (Lenski 1982, Lemieux and Lindgren 2004). Us ing this spec ies as a focus, I addressed the following quest ions: Do riparian reserves retain forest-preferring insects such as Scaphinotus angusticollis, to the s a m e extent as contiguous forest and do S. angusticollis distribute themse lves ac ross habitat in riparian buffers the s a m e way as in cont iguous forest? Pr ior knowledge about S . angusticollis al lowed me to hypothesize that riparian buffers may retain populat ion s i zes that are similar to that of cont iguous forest, because microcl imatic condit ions are similar enough in this habitat. However, clearcut riparian areas should have smal ler populat ions than either buffer or control (uncut) si tes. I a lso hypothesized that within riparian buffers, S . angusticollis would be found c loser to the streams, as this part of the habitat is best protected from edge effects of the highly var iable cl imate of c learcuts. Behavioural aspects of S. angusticollis' ecology such as movement are a lso important to examine because they represent the beetles' response to their environment and may indicate dif ferences in the perceived condit ions (Szyszko et al. 2004). Whi le it is difficult to conc lude that certain movement behaviours indicate less favourable habitat, I can address the more general quest ion: Does movement behaviour of S . 36 angusticollis differ between riparian buffers and cont iguous forest? B a s e d on prior studies on the movements of forest-preferring carabids in agricultural l andscapes (Rieken and Raths 1996), I hypothesize that S. angusticollis movement will be greater in the more open habitat of c learcuts, but that movement behaviour in buffers will not differ from cont iguous forest habitat. Methods Nine study sites for this exper iment were establ ished along nine s t reams in the University of British Co lumbia 's Ma lco lm Knapp R e s e a r c h Forest in Map le R idge, B C within forested areas c lose to smal l s t reams (<1m in width, max imum wetted width 2.4 m) in the riparian zone (Kiffney et al. 2000). Two different harvesting treatments (habitats) were se lected: clearcut to the edge of the s t ream (no buffers) and reserves with 30 m of intact forest remaining on both s ides of the st ream (30 m reserve). S i tes in areas with no harvesting were used as controls for compar ison . Three geographical ly-spaced repl icates of each treatment, nine st reams in total, were used . Al l nine st reams were located in the southeast corner of the Malco lm Knapp R e s e a r c h Forest , an area approximately 10 k m 2 . The southern area of Ma lco lm Knapp R e s e a r c h Forest is predominately western hemlock (Tsuga heterophylla), western red cedar (Thuja plicata) and Douglas fir (Pseudotsuga menziesii). The average stand age is 80 years old, with a history of stand-replacing fires. Forest harvesting in six of these nine sites occurred two years prior to this study (Kiffney et al. 2003). Gr ids of 100 pitfall traps that did not contain a killing agent were laid out to sample an area 45 m x 45 m immediately adjacent (<1 m) to the s t reams at each site; traps were equal ly spaced at 5 m intervals in ten rows and ten co lumns. Pitfall t raps 37 consisted of two plastic cups (diameter - 1 0 cm) embedded in the ground inside one another so that the lips of the cups were at ground level. Four p ieces of wood were laid in an x-pattern, extending 10 c m out from the edge of the cup, to guide walking insects into the trap. To keep rain water out of the traps, a 30 c m x 30 c m piece of plywood was laid on the four p ieces wood (Morrill et al. 1990). Gr ids of traps in 30 m reserve sites extended from the st ream's edge out into the clearcut adjacent to the reserve, with at least 30 traps located in the clearcut. Scaphinotus angusticollis were live trapped from mid-July 2000 to mid-Sep tember 2000 and again from mid-May 2001 to late Augus t 2001 . Live trapping in 2002 used different methods, a s descr ibed below. Al though other spec ies of large carabid beet les were captured in some of these pitfall t raps, too few were captured to al low for subsequent analys is . In 2000, trapping w a s done at five si tes in one week and four the next week, al lowing for a two-week rotation through all the si tes. Between trapping sess ions , traps were "deact ivated" by inserting st icks into the cups, so that animals falling into the cups could escape . E a c h week of trapping samp led habitats for four nights. In 2001 , three sites were trapped per week, al lowing for a three-week rotation through all si tes. At least one replicate of e a c h habitat type was sampled in each trapping sess ion , to minimize variation in results owing to seasona l activity or other temporal effects (e.g. weather affecting activity levels). N e w individuals trapped overnight in pitfall trap arrays were marked with a unique, three-digit number hand-painted with a paint pen on the elytra and weighed using an electronic field sca le (Ohaus Scout electronic field sca le , cal ibrated to 1.00 g). Subsequent ly , their pronotum was measured with cal ipers and they were re leased. The identity of e a c h recaptured 38 individual w a s recorded, and location and body measurements (e.g. weight, pronotum width) were retaken. Al though est imates of the population s i zes could be made using the mark-recapture data, only catch-per-unit-effort population est imates for 2000 and 2001 data were used in this analys is . Prel iminary analys is (Chapter Two) showed no significant di f ferences between the mark-recapture and the catch-per-unit-effort population est imates. A split-plot A N O V A was used to account for repeated sampl ing within the s a m e populat ion, with month a s the subplot factor. Statist ical analys is of C P U E est imates examined habitat, the nest ing of site in habitat as a random factor, month, the interaction of date by habitat and year. Y e a r was included in the model separately from month of capture to test for yearly variation in populat ion s izes . I selected these interactions because there were not iceable di f ferences in the raw data in 2001 (Figure 3.1). Prior to testing, data were found to not be normally distributed and were therefore log normal transformed prior to testing by A N O V A . A more detai led examinat ion of population distributions w a s given by calculat ing the mean number of individuals caught per month at each d is tance from the stream for each site. T h e s e data were only avai lable for 2000 and 2001 data and are ana lyzed and cons idered separately. Prior to statistical testing, the data were found to not be normally distributed, even after transformation. Rank data were therefore used in analys is . A split-plot A N O V A was used to ana lyze these data, with the month of capture used as the subplot factor, to account for repeated sampl ing within the s a m e populat ion. Statist ical ana lyses of populat ion distributions examined month, habitat, habitat c rossed 39 with month and d is tance from stream (row) nested within habitat as a random factor. T h e s e interactions were also selected for analys is because large dif ferences in population est imates between years were noted for 30 m buffer habitats. In 2002, the s a m e sites were used to catch and remove individuals from the forest for use in d issect ions for a s s a y of individual condit ion. The number of individuals caught per night in each site w a s converted to a catch-per-unit-effort est imate of populat ion s ize . Simi lar to the mark-recapture study, one replicate of each habitat type w a s visited during each trapping sess ion and all si tes were visited approximately once a month from May through to October. T h e s e population est imates were used in direct compar ison with mark-recapture data in this analys is. Movement analys is Movements of beet les in the mark-recapture study were ana lyzed for three different aspec ts : d istance travel led, movement with respect to riparian buffers, and directionality of movement . E a c h of these aspec ts of movement w a s selected to illustrate a different quality of movement and will be d i scussed separately. Of the recaptures made, only those that were made over 24 hours were ana lyzed so that a distinct t ime period could be ass igned to that d istance. To prevent multiple captures of one individual from biasing the data, only one overnight movement per individual was used . Owing to a low number of recaptures, no mean distance travelled could be calculated for clearcut habitat in July 2000 and August 2001 , or for control habitat in Ju ly 2001 . Prior to statistical testing, mean d is tances travelled overnight for each site and month tested for normality and heteroscedasci ty. E v e n after log transforming the data, 40 they were found to be non-normal , therefore ana lyses were done using rank data. Owing to the low number of recaptures during s o m e time periods, it w a s not possib le to run a detai led A N O V A model on these data; a one-way A N O V A w a s run using habitat as the main effect. Further examinat ion of d is tances travelled overnight was made to examine for possible effects of forested versus unforested a reas within 30 m reserves. For this analys is , all movement, regardless of habitat type (clearcut, 30 m reserve and control), w a s classi f ied as occurr ing within 30 m of the stream (close) or beyond 30 m from the st ream (far). A n y movements that t raversed this boundary were disregarded for this portion of the movement analys is . The A N O V A model looked at habitat as a separate factor, but site and "c lose or far" were nested with habitat type in the analys is . The last aspect of movement that was ana lyzed was the direction of movement. The direct ions the beetles travelled were broken down into x and y vectors (based on grid orientation) and the absolute value of these vectors (to avoid negative values) were natural log-transformed (Gardner et al. 1989) for this analys is. Whi le this treatment of the data does not allow for directionality of movement to be identified (e.g. towards the stream), it does identify whether x or y vectors in general were larger in some habitat types. A more detai led analys is of beetle movement vectors w a s not possib le because few recaptures were made in clearcut habitats and gave insufficient degrees of f reedom for models in S A S . Data analys is Al l ana lyses for this study were conducted in J M P (Vers ion 4.04; Cary , N C ) . Direct ional vectors were ana lyzed using M A N O V A (dependent var iables were x and y vectors) in S A S . 41 Resu l ts In total, 666 captures (including 64 recaptures) of S . angusticollis were made in 2000, 950 captures (including 96 recaptures) were made in 2001 , and 325 captures were made in 2002. Habitat types with the highest capture numbers varied from year to year and from season to s e a s o n , with control and 30 m buffer si tes having the most captures in all periods samp led , as shown by the least squared means (Figure 3.1). There w a s no overlap between 30 m buffer and clearcut si tes, but control sites spanned this gap between the two logged habitats (Figure 3.1). Variabil i ty in capture rates w a s general ly much higher in the last trapping sess ion of each year, but was consistent ly higher in 30 m buffer habitats than the other habitat types (Figure 3.1). There were significant effects of site nested within habitat (p = 0.01) and month (p = 0.02) on populat ion est imates (Table 3.1) Numbers of beet les in the three different habitats for 2000 and 2001 did not show any obvious linear pattern of beetle distribution with respect to the st reams in the three different habitats (Figures 3.2A and 3.2B). However, beetle distributions appeared to fol low a similar but weak pattern with peaks in abundance occurr ing at 15 and 45 m from the stream in 2000 (Figure 3.2A) and at 0 and 25 m in 2001 (Figure 3.2B). Relat ive abundances of beet les were not significantly lower in clearcut a reas of 30 m buffer si tes, because beetle distributions in the clearcut portions (at 35 , 40 and 45 m away from the stream) are at the same abundance levels as the forested area of the trapping grid (Figures 3.2A and B). In agreement with trends from the monthly C P U E populat ion est imates, the statistical analys is of the population distributions indicated there were significant effects of month (p = <0.0001) and some sites within habitats (p = 42 <0.0001) on populat ions as well a s the interaction of habitat and month (p = 0.01) (Table 3.2). T h e distance from the stream (row) a lso had a significant effect on population s i zes (p = 0.03) (Table 3.2). Movement rates in areas less than versus greater than 30 m from the stream showed that within some habitat types, the distance from the stream inf luenced the amount of movement by beet les (p = 0.07). Scaphinotus angusticollis movement at varying d is tances from the st reams were also significantly affected by s o m e sites within habitat types (p = 0.04) (Table 3.3). Scrut iny of the least square means for these effects shows that in c learcuts, movement within 30 m of the stream is almost 10 x higher than at d is tances greater than 30 m from the stream. In the second analys is of movement data, the mean distance travelled overnight varied from 13.6 m in 30 m buffers during Sep tember 2000 to 0 m in c learcuts for July of both years (Figure 3.3). Al though least square means show that movement was higher in 30 m buffers, it was not significantly higher than in no buffers (p = 0.08) (Figure 3.3). The number of recaptures in each site for each time period (Table 3.4) g ives the N values for each mean in Figure 3.3. The model used habitat type, owing to a low number of data points avai lable for s o m e time periods (Table 3.5). No significant difference w a s found in movement between habitat types (Table 3.5). M e a n movement vectors a lso showed considerable over lap between treatments (Figure 3.4), resulting in no significant dif ferences for either x or y vectors as tested by a M A N O V A ( p = 0.12). 43 Discuss ion Ripar ian reserves are commonly recommended as a forest management practice, with hopes that these remnants will provide not only protection for aquatic ecosys tems from harvest ing, but a lso that the forest will provide habitat for terrestrial spec ies that are unable to persist in clearcuts. For many spec ies that prefer forests, the benefits of riparian reserves have not been tested and the responses of forest-dwell ing spec ies of insects are particularly poorly understood. If regenerating forests are to be co lonized by forest-preferring spec ies , it is vital that the landscape include forest remnants like riparian reserves that can sustain forest-preferring spec ies and provide colonists. In this study, forested reserves near s t reams did retain populat ions of S . angusticollis that were similar in s ize to those of intact forest, even several years after logging had occurred. G iven the suitability of carabids as indicators of habitat d isturbance to forest insects (Rainio and N iemela 2002), these f indings suggest that populat ions of some forest-preferring insects may be retained in riparian buffers at levels similar to cont iguous forest, giving added value to this management practice. Whi le the use of forested buffers has been shown for vertebrates (e.g. Hannon et al. 2002) and for some moths (Haddad etal. 2003 , Monkkonen and Mutanen 2003, Varkonyi et al. 2003), there are few studies to indicate that forest fragments such as riparian buffers would retain populat ions of forest-dwell ing insects. There are qualit ies of riparian buffers that may make them particularly well-suited to providing refugia for forest-dwell ing spec ies . In riparian areas, s t reams alter the humidity and temperature of the terrestrial habitat adjacent to them, resulting in different 44 plant and animal communi t ies (Naiman and D e c a m p s 1997, Moore et al. 2005). Many studies on carabid beet les have invoked assoc ia t ions with microcl imates a s the determining factor in beet les ' habitat choice (e.g. Thiele 1977, S p e n c e and N iemela 1994). If riparian buffers have microcl imate condit ions which approximate cont iguous forest, there should be no dif ferences in the distributions of beetle populat ions with respect to st reams. Th is study clearly indicates that distributions of S . angusticollis across the landscape depend on the location of s t reams. There were similar patterns across the three habitat types surveyed, with a peak in abundance occurr ing at either 15 m or 25 m from the st ream (Figure 3.2). The area surveyed in 30 m buffer si tes included both clearcut and forested habitat, whereas clearcut and control s i tes were more homogeneous in their coarse habitat structure (i.e. either completely harvested or not). Despite this extreme difference in habitats from one side of the 30 m buffer si tes to the other, there w a s no significant drop in the number of beet les caught in the clearcut part of the 30 m reserve habitat compared to the forested side. Another finding of this study w a s that riparian buffers had a greater effect on the distribution of populat ions in one of the three years studied. Scaphinotus angusticollis population s izes in 30 m buffers were much larger in 2001 than they were in other habitats or years. O n e possib le explanat ion for this difference between years is that the clearcut habitats were a more inhospitable habitat in 2001 . Weather records for the Malco lm Knapp Resea rch Forest indicate that 2001 was hotter and drier than 2000 (Kiffney et al. 2002). Whi le local distributions of S . angusticollis did not appear to be affected by microcl imate changes from forest to clearcut in 30 m buffer habitat, it is 45 possib le that larger-scale distributions of S . angusticollis are dictated by f luctuations in yearly weather patterns. The number of S . angusticollis caught in riparian reserves w a s similar to or higher than those caught in cont iguous forest, however, p resence of a spec ies does not necessar i ly ensure residency in that area. It is not possib le to differentiate between individuals that are travelling through a habitat (dispersers and migrators) versus those who are using it for foraging (residents) without more detai led trapping data and behavioural analys is . Al though it is not possib le to differentiate between the uses of riparian reserves by S. angusticollis, the number of individuals caught in this type of habitat and the d is tances moved in reserves indicate that these individuals readily use this forest for either purpose. In addit ion, the clearcut area surrounding reserves appears to be used for movement in a similar manner to the adjacent forest within the reserve. In si tes located entirely in clearcut habitat, the movement pattern w a s entirely different, however, with the presence of s t reams increasing movement dramatical ly. The importance of movement proximal to reserves may emphas ize their role as a d ispersal corridor. Recent behavioural models of organ ism movement through corridors (Hudgens and Haddad 2005) indicate that where movement through the surrounding matrix (harvested clearcut) is high, corridor movement is unlikely to have much of an effect on populat ions. A s shown by results from grids entirely located in the clearcut, movement rate was lower, al though not significantly. Movement by invertebrates through riparian buffers has received little attention until recent years. Studies on the movement of moths (Monkkonen and Mutanen 2003 , Varkonyi et al. 2003) suggest that riparian buffers may be used as habitat by forest-46 preferring spec ies and for d ispersal between forest f ragments. However, the d ispersal ability of moths is higher than most ground-dwell ing insects (Haddad 1999, Monkkonen and Mutanen 2003, Varkonyi et al. 2003) and habitat requirements are a lso quite different. The results of our study showed that a similar amount of movement occurred within the 30 m riparian buffers when compared to control si tes and suggests that this habitat could be used for d ispersal between forest f ragments by f l ightless insects. Individuals in clearcut si tes did not show the s a m e level of mobility at a d is tance from st reams and it is possible that this habitat is more difficult for them to move in. However, smal ler movement d is tances could a lso be attributable to other factors such as smal ler home range s izes and different foraging behaviour (Baars 1979) that cannot be tested using the data col lected in this study. Regard less of the interpretation of the amount of movement, permeabil i ty of riparian buffers to forest-dwell ing insects is a critical feature of this habitat, if it is to serve any purpose as a d ispersal corridor between areas of cont iguous forest. Whi le riparian buffers were originally intended to provide protection for s t reams, there are evidently some benefits to forest-dwell ing insects as wel l . Pas t studies on S . angusticollis have detected that smal l patches of forest can harbour this spec ies both immediately after harvest (Laval lee 1999) and for severa l years post-harvest (Lemieux and Lindgren 2004), but the linear structure of riparian buffers has not specif ical ly been tested. In addit ion, these studies removed large numbers of individuals from the population using catch-kil l methods that may bias results of long-term studies, especia l ly when working with a spec ies that may live severa l years in its adult form (see Chapter Four). Whi le these riparian buffers offer habitat that appears to be adequate to 47 support populations of forest-dwell ing beet les such as S. angusticollis, caution should be used in equating buffers with cont iguous forest. Large populat ion s i zes in pass ive sampl ing studies may indicate large populat ions and/or high amounts of movement in the population. In addition to this difficulty in interpretation, large amounts of movement in a population may be s ignals of large migration, good foraging, or stressful habitats (Baars 1979). More detai led behavioural studies of S . angusticollis movement and body condit ions would clarify how to interpret these responses to different habitats. 48 Literature Ci ted Baars , M.A. 1979. Patterns of movement of radioactive carabid beetles. Oeco log ia 44: 125-140. Cock le , K.L. and J . S . R ichardson . 2003 . Do riparian buffer strips mitigate the impacts of clearcutt ing on smal l m a m m a l s ? Biological Conservat ion 113: 133-140. Darveau, M., P. Beauchesne , L. Belanger , J . Huot, and P. Larue. 1995. Ripar ian forest strips as habitat for breeding birds in boreal forest. Journa l of Wildl i fe Management 59: 67-78. Elphick, C . S . 2004. A s s e s s i n g conservat ion trade-offs: identifying the effects of f looding rice f ields for waterbirds on non-target bird spec ies . Biological Conservat ion 117: 105-110. Gardner , R .H . , O'Nei l l , R.V. , Turner, M . G . and V . H . Dale . 1989. Quantifying sca le -dependent effects of animal movement with s imple percolat ion models . Landscape Eco logy 3: 217-227. Haddad , N .M. 1999. Corr idor length and patch colonizat ion by a butterfly, Junonia coenia. Conservat ion Biology 14: 738-745. Haddad , N .M. , D.R. Bowne, A . Cunn ingham, B . J . Dan ie lson , D .J . Levey, S . Sargent , and T. Sp i ra . 2003 . Corr idor use by diverse taxa. Eco logy 84: 609-615. Hannon, S . J . , C A . Paszkowsk i , S . Bout in, J . DeGroot , S . E . Macdona ld , M. Wheat ley, and B.R. Eaton . 2002. Abundance and spec ies composi t ion of amphib ians, smal l mammals , and songbirds in riparian forest buffer strips of varying widths in the boreal mixedwood of Alberta. Canad ian Journa l of Forest Resea rch 32:1784-1800. Hannon S . J . and F .K.A. Schmiege low. 2002. Corr idors may not improve the conservat ion value of smal l reserves for most boreal birds. Eco log ica l Appl icat ions 12: 1457-1468. Hi lderbrand, R .H . , A . C . Wat ts and A . M . Rand le . 2005. The myths of restoration ecology. Eco logy and Soc ie ty 10: 19-28. Hudgens , B .R. and N .M. Haddad . 2003 . Predict ing which spec ies will benefit from corridors in f ragmented landscapes from population growth mode ls . The Amer i can Naturalist 161: 808-820. Hylander, K., C . N i lsson, and T. Gothner . 2004. Effects of buffer-strip retention and clearcutt ing on land snai ls in boreal riparian forests. Conservat ion Biology 18: 1052-1062. Kiffney, P . M . , J . S . R ichardson , and M .C . Fel ler. 2000. Fluvial and epilithic organic matter dynamics in headwater s t reams of southwestern British Co lumb ia , C a n a d a . Archiv fur Hydrobiologie 149 :109 -129 . Kiffney, P . M . , J . P . Bul l , and M . C . Fel ler. 2002. Cl imat ic and hydrologic variability in a coasta l watershed of southwestern British Co lumb ia . Journal of the Amer i can Water R e s o u r c e s Assoc ia t ion 8: 1437-1451. Kiffney, P . M . , J . S . R ichardson , and J . P . Bul l . 2003 . R e s p o n s e s of periphyton and insects to manipulat ion of riparian buffer width a long forest s t reams. Journal of Appl ied Eco logy 40: 1060 Koivu la , M. , J . Kukkonen, and J . N iemela . 2002. Borea l carabid beetle (Coleoptera, Carab idae) assemb lages along the clear-cut originated success ion gradient. Biodiversity and Conservat ion 11: 1269-1288. 50 Laval lee, S . L . 1999. C h a n g e s in the carabid beetle communi ty (O. Co leoptera , F. Carab idae) of the S i c a m o u s Creek R e s e a r c h Si te. Master 's thesis, Dept. of Zoo logy, U B C . P p . i to 73. L e e , P. , C . Smyth , and S . Boutin. 2004. Quantitative review of riparian buffer width guidel ines from C a n a d a and the United States. Journal of Environmental Management 70: 165-180. Lemieux, J . P . and B . S . L indgren. 2004. Ground beetle responses to patch retention harvest ing in high elevation forests of British Co lumb ia . Ecography 27: 557-565. Lensk i , R . E . 1982. The impact of forest cutting on the diversity of ground beetles (Coleoptera: Carab idae) in the southern Appa lach ians . Ecolog ica l Entomology 7: 385-390. Lindroth, C . H . 1961. The Ground Beet les of C a n a d a and A l a s k a . Opuscu la Entomologica Supp lementum X X - X X I V . P p . 1-1192. Mat lack, G . R . and J . Monde . 2004. C o n s e q u e n c e s of low mobility in spatially and temporal ly heterogeneous ecosys tems. Journal of Eco logy 92: 1 0 2 5 - 1 0 3 5 Monkkonen , M. and M. Mutanen. 2003 . Occur rence of Moths in Boreal Forest Corr idors. Conservat ion Bio logy 17: 468-475. Moore , R.D. , D.L. Spi t t lehouse, and A . Storey. 2005. Ripar ian microcl imate and st ream temperature response to forest harvest ing: a review. Journa l of the Amer i can Water Resou rces Assoc ia t ion 41 : 813-834. Morril l, W .L . , D .G . Lester, and A . E . Wrona . 1990. Factors affecting eff icacy of pitfall traps for beet les (Coleoptera, Carab idae and Tenebr ionidae) . Journal of Entomologica l S c i e n c e 25: 284-293. 51 Naiman R . J . and H. Decamps . 1997. The ecology of interfaces: Ripar ian zones . Annua l Rev iew of Eco logy and Systemat ics 28: 621-658. Potv in, F., N. Bertrand, and J . Ferron. 2005. Attributes of forest strips used by snowshoe hare in winter within clear-cut boreal landscapes. Canad ian Journal of Forest R e s e a r c h 35: 2521-2527. Rain io , J . and J . N iemela . 2002. Ground beetles (Coleoptera: Carab idae) as bioindicators. Biodiversity and Conservat ion 12: 487-506. R ichardson , J . S . , R . J . Na iman, F . J . S w a n s o n and D.E. Hibbs. 2005. Ripar ian communi t ies assoc ia ted with Paci f ic Northwest headwater s t reams: assemb lages , p rocesses , and un iqueness. Journal of the Amer ican Wate r R e s o u r c e s Assoc ia t ion 41 : 935-947. R iecken , U. and U. Raths. 1996. U s e of radio telemetry for studying d ispersal and habitat use of Carabus coriaceus L. Annates Zoologic i Fennic i 33 : 109-116. Shir ley, S . M . and J . N . M . Smith. 2005. Bird community structure ac ross riparian buffer strips of varying width in a coasta l temperate forest. Biological Conservat ion 125: 475-489. S p e n c e , J . R . and N iemela , J . K . 1994. Sampl ing carabid assemb lages with pitfall traps: the madness and the method. Canad ian Entomologist 126: 881-894. S w a n s o n , F . J . and J . F . Frankl in. 1991. New forestry principles from ecosys tem analys is of Paci f ic northwest forests. Eco log ica l Appl icat ions 2: 262-274. S z y s z k o , J . , S . Gryuntal , and A . Schwerk . 2004. Dif ferences in locomotory activity between male and female Carabus hortensis (Coleoptera: Carab idae) in a pine 52 forest and a beech forest in relation to feeding state. Environmental Entomology 33 : 1442-1446. Thie le, H .U. 1977. Carab id Beet les in their Environments. Spr inger Ver lag , Berl in. P p . xvi i -369. Varkony i , G . , M. Kuussaar i , and H. Lappala inen. 2003. U s e of forest corridors by boreal Xestia moths. Oeco log ia 137:466-474. Y o u n g , K.A. 2000. Ripar ian zone management in the Paci f ic Northwest: W h o ' s cutting what? Environmental Management 26: 131-144. 53 Tab les Table 3.1 Analysis of variance of catch per unit effort (CPUE) for Scaphinotus angusticollis in forest, 30 m buffer, and clearcut sites. Main effects include habitat type (H), Site (H), month, the interaction of month and habitat and year. (Model r2 = 0.49, df = 16) Source df S S MS F P Habitat (H) 2 31.44 15.72 0.52 0.62 S ( H ) 6 195.18 32.53 3.17 0.01 Month (M) 2 94.54 47.27 4.60 0.02 M*H 4 34.71 8.68 0.85 0.50 Year 2 21.81 10.91 1.06 0.36 Error 40 410.61 10.27 Total 56 806.34 Table 3.2 Analysis of variance of catch per unit effort population estimates for Scaphinotus angusticollis in forest, 30 m reserve, and clearcut sites at varying distances from the stream. Main effects include habitat type (H), month (M), row (R), the nested term of site within habitat type, and the interactions of M*R, H*M, H*R, and H*M*R. (Model r2 = 0.64, df =155) Source DF Sum of Squares MS F P Habitat (H) 2 26.59 13.29 0.62 0.57 Row 9 14.80 1.64 2.12 0.03 Month (M) 4 97.34 24.33 31.40 <0001 Site (H) 6 133.27 22.21 28.66 <0001 M * R 36 32.14 0.89 1.15 0.26 H * M 8 15.60 1.95 2.52 0.01 H •* R 18 16.48 0.92 1.18 0.27 H * R * M 72 40.80 0.57 0.73 0.94 Error 293 227.06 0.77 Total 448 267.25 54 Table 3.3 Analysis of variance for distances moved by Scaphinotus angusticollis in clearcut, 30 m reserve and control sites at distances less than 30 m from the stream (close) and greater than 30 m from the stream (far). Main effects include habitat (H), site nested within habitat, and "close or far" nested within habitat. (Model r2 = 0.94, df = 10) Source DF Sum of Squares F Ratio P Habitat (H) 2 17.94 2.98 0.16 Site (H) 5 106.37 7.08 0.04 "Close or far" (H) 3 50.03 5.55 0.07 Error 4 12.03 3.01 Total 14 190.80 Table 3.4 Total number of recaptures of S. angusticollis made in control, 30 m buffer and clearcut sites in 2000 and 2001. Site names are provided under each habitat type. Habitat fype-> Control 30 m buffer Clearcu t S i t e ^ G Mike Moss Total D H South Total I E B Total Date July 2000 11 3 0 13 1 1 1 3 0 0 0 0 • I B Aug 2000 5 7 0 12 1 2 5 8 0 1 1 2 Sept 2000 3 4 0 7 0 0 4 4 3 0 11 11 May/June 2001 0 0 . 6 6 2 0 16 18 0 0 1 1 July 2001 0 0 0 0 4 0 35 39 0 0 0 0 Aug 2001 4 3 0 i l l l l l 1 3 16 20 0 4 0 4 Table 3.5 Analysis of variance of distances moved by Scaphinotus angusticollis in clearcut, 30 m reserve and control sites. Main effect is habitat only. (Model r2 = 0.07, df = 2) Source DF Sum of Squares MS F Ratio Habitat 2 7.25 0.28 0.76 Error 8 102.75 12.84 Total 10 110.00 55 Figures • Control 30 m buffer • Clearcut 2000 / / / 2001 2002 Figure 3.1 Catch-per-unit-effort (CPUE) population estimates (not including any recaptures) of S. angusticollis for control (circle), 30 m buffer (triangle) and clearcut (square) sites in 2000, 2001 and 2002. Error bars indicate standard error. N=3 for all trapping dates except N=2 September/October 2000 and 2002. Asterisk (*) indicates significant difference (p = 0.03). 56 3.0 2.5 2.0 A "1 1.5 1.0 0.5 0.0 • Control W 30 m buffer • Clearcut A —i 1 ~i 1 1 1 1 1 1 1 — 0 5 10 15 20 25 30 35 40 45 50 10 15 20 25 30 35 Distance from stream (m) 50 Figure 3.2 Distribution of mean numbers of S. angusticollis caught per trap night in control, 30 m buffer and clearcut habitats in 2000 (A) and 2001 (B) at varying distances from the stream. Dotted line indicates forest boundary in 30 m buffer sites. Error bars represent standard error. (N (number of trapping sessions) = 11 (control 2000) 12 (30 m buffer 2000) 11 (clearcut 2000) and N = 14 (all habitats 2001)) 5 7 o V H Control 30 m reserve Clearcut i • 0 • May/June July August Date September Least Square Means Figure 3.3 Mean distance moved overnight by S. angusticollis in control (circle), 30 m buffer (triangle), and clearcut (square) habitat in 2000 (black) and 2001 (grey). Error bars represent standard error. See Table 3.3 for N of each point. o.o 4-• Control f 30 m buffer • Clearcut -1.0 -0.5 0.0 0.5 1.0 1.5 Ln Mean horizontal distance (m) 2.0 2.5 Figure 3.4 Mean movement vectors (In) for distances travelled overnight by S. angusticollis in control, 30 m buffer and clearcut sites in 2000 and 2001 (pooled). Error bars represent standard error. N= 7 (clearcut), 26 (30 m buffer), and 15 (control), not weighted by site within habitat. 5 8 Chapter Four Snail (Gastropoda, Pulmonata) abundance and diversity in three forest habitats of Pacific Northwest coniferous forests (Canada) Introduction Terrestrial snail conservat ion has become a topic of greater concern with the recognit ion of snail sensitivity to habitat d isturbance (Hylander et al. 2004). Sna i l s and s lugs are abundant and voracious detritivores in many forest ecosys tems , convert ing plant matter to a partially-digested form for other plants and organ isms (Forsyth 2004). In addit ion, many snai ls prey on each other, consuming eggs and shel ls of other gastropods (Desbuquois and M a d e c 1998). Whi le their distributions are in part determined by soil types (Cameron 1986) and ca lc ium sources (Wareborn 1969, Pr ior 1985), they rely heavily on a steady supply of herbaceous material from the vegetat ion. B e c a u s e of their sensitivity to des iccat ion and their inability to d isperse long d is tances, forest pract ices are of particular concern to snai l conservat ion, where smal l sca le management may have great effects on gast ropods (Hylander et al. 2004). Forest harvesting, an important and widespread form of habitat d isturbance, alters many of the p rocesses that forests depend on . Decomposi t ion and the suite of organisms that provide the convers ion of organic matter are often over looked by studies that examine productivity or invertebrate responses to harvest ing. B e c a u s e they are such a poorly recognized and understood group, gastropods are rarely included in studies of forest ecology (Suominen et al. 2003). Predict ions regarding the influence of logging on snai l populat ions are difficult to make , because previous studies have indicated both increases and dec reases in snai l abundance after logging (Hylander et al. 2004). The new abiotic condit ions present in a logged habitat may favour certain 59 spec ies over others (Schowal ter et al. 2003), depending on their ability to use refugia from extreme microcl imatic condit ions (Hylander et al. 2004) or to withstand desiccat ion (Prior 1985). Sna i ls make good study subjects for field surveys, as even dead spec imens are often identifiable to the genus level and they are relatively e a s y to find when searching by hand. Despite this, few studies have examined gastropod responses to habitat change , possibly because of the large amount of t ime required to complete hand sea rches and the lack of attention given to this taxon in eco log ica l studies (Hylander et al. 2004). Al though there has been some work done to examine snai l distributions within and around boreal forest habitats (e.g. Hylander et al. 2004, 2005) and some examinat ion of snai l assoc ia t ions with soi l types in Paci f ic Northwest forests (Cameron 1986), there has been little examinat ion of the effects of forest harvesting on snai l diversity in temperate forests. More locally, natural history studies (Branson 1977) and surveys of local forested habitats (Cameron 1986) provide a bas is for this study. In this survey of snai l populat ions, I sought to examine the diversity and abundance of large (>10 mm), pulmonate snai l spec ies in three types of harvested habitat: clearcut, partially harvested, and uncut si tes. Th is study w a s originally des igned and planned to supplement studies on Scaphinotus angusticollis (Chapters Three and Five), which has been identified by some researchers as a special ist predator of snai ls (Thiele 1977, Digweed 1993). Al though s lugs are a lso common in Paci f ic Northwestern forests, exper imental feeding studies by D igweed (1993) showed that s lugs are consumed when cut into smal ler p ieces, which is unlikely to occur in nature. B e c a u s e of the higher b iomass of herbaceous plant material in c learcuts and 60 based on prior studies (Cameron 1986, Hylander etal. 2004), I hypothesized that the overal l abundance of snai ls would increase in clearcut habitat. The diversity of snai ls present was a lso expected to be higher within c learcuts, because the amount of food material for snai ls is clearly more abundant in habitat with more herbaceous growth. Methods Three habitat types were se lec ted for this study: clearcut, 30 m riparian forest reserves, and uncut, second-growth forest. Three geographica l ly -spaced repl icates of each forest type were used . In the two habitats with logging, harvesting w a s done four years prior to this study. T h e s e nine study sites were part of a larger study on the management of forests proximal to smal l s t reams; (Kiffney et al. 2003). Hand sea rches for snai ls were conducted at ten stat ions separated by 5 m along a 45 m long transect line that extended perpendicularly out from the st ream. S e a r c h e s involved light d isturbance of the soil by hand to a depth of 3 c m and turning over of any large wood found within a i m radius from the station. The bases of ferns and any leaf litter were c losely examined for live snai ls or any snai l shel ls . Th is search technique w a s fashioned after recommendat ions by Hawkins et al. (1998). S lugs were not evaluated in this study. Sna i l s were categor ized in the field by morphospec ies , with voucher spec imens taken for more detai led identification. Identification of shel ls to genus w a s done using Branson (1977) and confirmation of identif ications (from spec imens extracted from the study sites) w a s made using Forsyth (2004). The number and location of each type of snai l was recorded at e a c h site from M a y to August in 2002. E a c h site w a s searched once in this time period. T h e s e data were ana lyzed using Krebs ' Eco log ica l 61 Methodology statistical package (1999) with Shannon -Weave r Diversity Index calculated for e a c h site. V isua l est imates of the percentage cover from vegetat ion, large downed wood (larger than 6 c m in d iameter and longer than 1 m) and s lash (fine wood p ieces less than 6 c m in diameter) were taken along the ten-station transect in each of the nine sites surveyed for snai ls. A n area of 1 m radius from the station w a s measured and then visual ly surveyed by two people. The mean of the two est imates was taken as the amount of cover. Est imates were tested against individual spec ies abundances and S h a n n o n - W e a v e r diversity va lues for each site using canonica l cor respondence analys is ( C C A ) in C A N O C O (Vers ion 4.0) to determine which of the three environmental var iables w a s the best descr iptor of individual snai l abundances and diversity. Canon ica l cor respondence analys is provides an advantage over l inear regression and canonica l correlation analys is , because there is no presumption that a l inear relationship is the best descr iptor between a spec ies and any given environmental variable (ter Braak and Smi lauer 1998). Detrended cor respondence analys is was inappropriate for these data, owing to the number of ze roes in the spec ies occurances , which a lso precluded analys is by mult idimensional scal ing (Manly 2005) . B e c a u s e it is difficult to predict the relationship of individual snai l spec ies with environmental var iables a priori, canonica l cor respondence analys is w a s se lected to analyze these data. In addit ion to this advantage, canonica l cor respondence analys is is better for data sets of spec ies occur rences where zero va lues are common . 62 Resu l ts Only three spec ies of snai ls were found in the surveys conducted for this study. Ancotrema hybridum A n c e y (F. Haplotrematidae), Cryptomastix germana Gou ld (F. Polygyr idae), and Vespericola columbianus I. L e a (F. Polygyr idae) were all found in e a c h habitat type, al though in greatly varying numbers within and between habitat types (from 1 to 9 individuals) (Figure 4.1). Us ing the hand search methods descr ibed above, smal ler spec ies of snai ls (< 5 mm) may have been over looked, but this low number of spec ies is an accurate representation of the large snail (> 5 mm) diversity in these habitats, as found by C a m e r o n (1986). The mean number of snai ls for all logging types I examined over lapped significantly (Figure 4.1). In the clearcuts, the mean number of A. hybridum w a s three t imes greater than in other habitats; but other spec ies were found in approximately the s a m e , low abundances in clearcuts (Figure 4.1). In 30 m buffer habitat there w a s three t imes the mean number of V. columbianus compared to the other two spec ies (Figure 4.1). The mean Shannon -Weave r diversity indices were very similar for the three habitats examined (Figure 4.2), ranging from 0.62 to 0.72. However, high variability in the est imates for certain si tes resulted in a large standard error for these means . Canon ica l cor respondence analys is results (Table 4.1) indicated that the first e igenvalue (0.33) expla ined less than half the overal l inertia with p = 0.22, indicating that these habitat components were a poorly-fitting model that expla ins little of the variation in the data. Resampl ing using Monte Car lo post hoc analys is indicated that the sum of all e igenvectors was 0.53 with overal l p = 0.39, indicating that little of the 63 data was expla ined in this model with low signi f icance. Al though none of these assoc ia t ions w a s significant, A. hybridum w a s best related to plant cover and fine woody debris, which is consistent with the higher abundance of this snai l in c learcuts (Figure 4.1). Vespericola columbianus was best related to coarse woody debris. However , C. germana was not well-related to any of the measured environmental var iables, possibly owing to the smal l number of individuals in the survey. D iscuss ion Ancotrema hybridum, Cryptomastix germana, and Vespericola columbianus are all commonly found in coniferous and mixed forest types in the Paci f ic Northwest, at a range of elevat ions (Forsyth 2004). Whi le C. germana is c losely related (Family Polygyr idae) to Cryptomastix devia, which is listed as extirpated, and Allogonia townsendiana, which is listed as endangered by C O S E W I C ( C D C Webs i te , March 2006; ht tp: / /www.env.gov.bc.ca/cdc/) , there is no current concern over the status of this snai l . Typical ly, C. germana occurs in low numbers (Forsyth 2004) and no studies have documented population decl ines or loss of spec ies range due to human disturbances. G iven the low numbers of C. germana observed in this study, no further asser t ions can be made regarding the impacts of forest harvesting on this spec ies . S o m e spec ies observed in this study appeared to have different assoc ia t ions with forest harvesting proximal to st reams. In past studies, total snai l abundance has been noted to increase 3 - 4 years after harvest ing, presumably owing to an increase in the herb vegetat ion, snai ls ' main food source (Cameron 1986). Th is study indicates that not all spec ies may have large population abundances after harvesting and suggests that condit ions in clearcuts may only be tolerable to some spec ies . The mean number 64 of snai ls w a s most variable in clearcut habitat, possibly because of the heterogeneity of the clearcuts used in this study. At the time of this study, s o m e sites had regenerated considerable ground cover from bracken fern [Pteridium aquilinum), a lder (Alnus rubra) and blackberry (Rubus sp.) (average vegetat ion cover est imate 58.5%), while others remained relatively bare (average vegetation cover est imate 19.5%). Prior studies (Cameron 1986) suggest that the herb layer is vital to pulmonate snai l populat ions; however, only A. hybridum abundance was best descr ibed by plant cover in C C A . Large downed wood may play an important role in retaining snai ls such as V. columbianus in harvested areas . Al though typically found in forested sites in this study, V. columbianus distributions were weakly correlated with coarse woody debris overal l . Anecdota l ev idence suggests that the unders ides of large downed wood were used as a refugium by snai ls during the day, because this w a s a common location to find snai ls in harvested and unharvested a reas (pers. obs.) . Th is finding is supported by Boag (1990) who suggested that masoni te boards could be used to survey for gastropods. The use of moist crev ices as refugia for land snai ls during harvest ing has also been suggested by Hylander et al. (2004). T h e s e studies add to the growing list of forest biota (e.g. vo les, saproxyl ic beet les, fungi) that rely on large downed wood in clearcuts, further emphas iz ing its importance in conservat ion of lesser-known, forest-dwell ing spec ies . Al though other aspects of large downed wood management and select ion such as s ize , decay c lass , and p lacement were not ana lyzed in this study, further research is needed to character ize the use of downed wood by pulmonate snai ls, in the interests of conserv ing snai ls in this habitat type. 65 Low numbers of pulmonate snai ls were not assoc ia ted with logging but the results of this study may have other implications for snai l conservat ion. S p e c i e s interactions between the numerical ly-dominant spec ies of snai ls and rarer snai ls may prevent re-establ ishment of some snail spec ies . Cann iba l i sm within the gastropods is not uncommon, and consumpt ion of other snai ls for their shel ls is noticeable in both live and dead spec imens (Branson 1977). G i ven that snai ls are typically limited by the amount of ca lc ium in their environment (Prior 1985), exploitation of such a concentrated source of nutrients is not surprising. The higher abundances of A. hybridum in c learcuts d o e s not appear to preclude other spec ies from occurring in clearcut habitat, which may suggest that competit ive interactions are not as important as other factors in determining spec ies distributions. The increase in a congener A . sportella was noted by Schowal ter et al. (2003) in thinned stands a s wel l . In order to explicitly test this hypothesis, evaluat ion of resource limitations and competit ive interactions would have to be conducted. Al though snail conservat ion does not receive as much attention as that for vertebrates or even insects (Hylander et al. 2004), conserv ing forest-dwell ing spec ies may not require large-scale reserves and may be combined with conservat ion of other spec ies . Moist a reas proximal to riparian areas may be protected owing to other constraints on harvest ing, such as those imposed for f ish-bearing st reams (Young 2000). However, gastropod conservat ion has a unique chal lenge in consider ing the poor dispersal capabil i t ies of snai ls and s lugs, which may move less than 5 m over 5 days (Hawkins et al. 1998). This study suggests that snai l surveys and conservat ion 66 efforts should cons ider individual spec ies responses and habitat needs, because overal l measures of abundance do not detect changes in individual spec ies . 67 Literature Ci ted B o a g , D.A. 1990. O n the effect iveness of artificial shelters on the study of population attributes of smal l terrestrial gastropods. Canad ian Journal of Zoo logy 68: 254-262. B ranson , B.A. 1977. Freshwater and Terrestrial Mol lusca of the Olympic Pen insu la , Wash ington. The Vel iger 19: 310-330. C a m e r o n , R .A .D . 1986. Environment and diversit ies of forest snai l faunas from coasta l British Co lumbia . Malacolog ia 27: 341-355. Desbuquo is , C . and L. Madec . 1998. Within-clutch egg canniba l ism variability in hatchl ings of the land snai l Helix aspersa (Pulmonata : Sty lommatophora): Influence of two proximate factors. Malaco log ia 39: 167-173. D igweed, S . C . 1993. Select ion of terrestrial gastropod prey by Cychr ine and Pterost ichine ground beetles (Coleoptera: Carab idae) . Canad ian Entomologist 125: 463-472. Forsyth, R . G . 2004. Land Sna i ls of British Co lumbia . Roya l B C M u s e u m , Victor ia, B C . Hawkins, J .W. , M.W. Lankester , and R .R .A . Ne lson . 1998. Sampl ing terrestrial gast ropods using cardboard sheets. Malacolog ia 39: 1-9. Hylander, K., C . Ni lsson, and T. Gothner. 2004. Effects of buffer-strip retention and clearcutt ing on land snai ls in boreal riparian forests. Conservat ion Biology 18: 1052-1062. Hylander, K. C . Ni lsson, B . G . J o n s s o n , and T. Gothner. 2005. Dif ferences in habitat quality explain nes tedness in a land snail meta-community. O ikos 108: 351-361. 68 Kiffney, P . M . , J . S . R ichardson, and J . P. Bul l . 2003. R e s p o n s e s of periphyton and insects to manipulation of riparian buffer width a long forest s t reams. Journa l of App l ied Eco logy 40: 1060-1076. Krebs , C . J . 1999. Ecolog ica l Methodology. Ben jamin /Cummings : Don Mil ls, O N . Manly, B . F . J . 2005. Multivariate Statist ical Methods: A Pr imer. C h a p m a n Hal l , N e w York. Prior, D .J . 1985. Water regulatory behaviour in terrestrial gast ropods. Bio logical Rev iews 60: 403-424. Schowal ter .T.D. , Y . L . Z h a n g , and J . J . Rykken . 2003 . Litter invertebrate responses to variable density thinning in western Wash ington forest. Eco log ica l Appl icat ions 13: 1204 -1211 . Suominen , O. , L. Eden ius , G . E r i csson , and V . R e s c o de Dios. 2003. Gas t ropod diversity in aspen stands in coastal northern S w e d e n . Forest Eco logy and Management 175: 403-412. ter Braak, C . J . F . and P. Smi lauer. 1998. C A N O C O Reference Manua l and User ' s Gu ide to C a n o c o for Windows (Vers ion 4). Centre for Biometry, Wagen ingen . Thie le, H .U. 1977. Carab id Beet les in their Environments. Spr inger Ver lag , Ber l in. P p . xvi i-369. Wareborn , I. 1969. Land mol luscs and their environments in an oligotrophic a rea in southern S w e d e n . Oikos 20: 461-479. Y o u n g , K.A. 2000. Ripar ian zone management in the Paci f ic Northwest: W h o ' s cutting what? Environmental Management 26: 131-144. 69 Tab les Table 4.1 Canonical correspondence analysis results comparing species occurrences with habitat features (plant cover, coarse woody debris, and fine woody debris). Axes Coarse wood cover Plant cover Habitat Fine wood (<6 cm) cover Total inertia Eigenvalues 0.330 0.052 0.077 0.066 0.526 Species-environment correlations 0.000 0.605 0.449 0.000 Cumulative percentage variance of species data 87.4 24.6 14.7 100.0 of species-environment relation 0.000 100.0 59.7 0.000 Sum of all unconstrained eigenvalues 0.526 Sum of all canonical eigenvalues 0.129 70 Figures Control 30 m buffer Habitat type Clearcut Figure 4.1 Mean number of snails found in control, 30m buffer and clearcut habitats along 45 m transects in 2002. N=3 for each habitat type. Error bars indicate standard error. CD > _ CD <T~ CD X X CD T3 C 1.4 1.2 1.0 4 0.8 ~ 0.6 4 0.4 0.2 H 0.0 Control 30 m buffer Habitat Type Clearcut Figure 4.2 Mean Shannon-Weaver Diversity Index (H1) values for control, 30m buffer and clearcut site snail diversity in 2003. N=3 for all habitat types, error bars indicate standard error. 71 Chapter Five Forest harvesting and the associated body conditions of the carabid beetle Scaphinotus angusticollis Introduction Environmental condit ions play a pivotal role in determining the viability of populat ions. The pers is tence of a spec ies integrally depends on food availability and abiotic factors such a s temperature and humidity. Envi ronmental st ressors may affect individuals and populat ions through short-term p rocesses such as changes in respiration rate (Bennett et al. 1999) or long-term p rocesses such as fecundity and population growth (Jul iano 1986). Phys io log ica l responses to st ress (e.g. re lease of glucocort icosteroids) have been used in recent studies to quantify stress in vertebrate populat ions and to predict the consequences of habitat alterations (Romero 2004). A l though this s a m e group of hormones is present in insects, they are related to diuresis and feeding (Lovejoy and J a h a n 2006). Ana logues of the hormones involved in vertebrate stress responses have not been isolated as of yet in insects (Lovejoy and J a h a n 2006), but the implications of s t ress are wel l -known. High st ress responses in insects have been linked to several body condit ion indicators such as growth, blood sugar levels and supress ion of reproduction a s wel l 'as behavioural changes and supress ion of immune de fences (Romero 2004). Whi le short-term stressors may not necessar i ly inf luence long-term responses (Moehrl in and Jul iano 1998), short-term responses such a s reduction in individual growth must be either overcome through compensatory growth (Dmitriew and R o w e 2005) or they will result in smal ler adult body s i zes (B lanckenhorn 2006) and/or reduced adult survivorship (Boggs and F reeman 2005). The types of environmental condit ions that pose physiological 72 stress on individuals will vary greatly in different ecosys tems and habitats and between different taxa. The strong ties between individual responses to habitat st ressors and population pers is tence has resulted in an emerging discipl ine that uses physiology and body metrics to reveal the mechan isms for population dec l ines, coined as "conservat ion physiology" by Wike lsk i and C o o k e (2006). Whi le directly observ ing physiological changes is not a lways feasible, the effects of past stress on individuals can be est imated using internal body condit ion as an indicator (Ricklefs and Wike lsk i 2002) Recen t work in agricultural ecosys tems (Ostman et al. 2001) has illustrated the use of internal body condit ions to evaluate different agricultural methods. Whi le the measurement of body condit ion in response to environmental factors has a long history in insect eco logy (Statzner et al. 2001), the use of body condit ions to identify potential mechanist ic responses of non-pest insects to different management strategies is uncommon. Forest harvest ing results in vast changes of microcl imatic condit ions (Chen etal. 1999) that must be overcome or avo ided by organ isms that prefer the more moderated environment of cont iguous forest. Ex t remes may be avoided through behavioural changes such as entering aestivat ion or physiological changes such as increased respiration (Bennett et al. 1999), that allow for the conservat ion of energy in physiological ly stressful environments. Another effect of increased temperatures is dehydrat ion, one of the most important physiological problems for insects in many environments (Todd and Block 1997). Average dayt ime temperatures in c learcuts are higher and humidity lower than in forests during the summer, but at nighttime, air 73 temperatures drop and humidity increases (Chen et al. 1999). For nocturnal insects, c learcuts may not be as dry as forest in the summer t ime because the lower temperatures al low dew to form. Food resources may also change after logging, if new resources in c learcuts are not access ib le to spec ies that depend on forest-dwell ing spec ies . Food is thought to be a limiting factor for many predators (Jul iano 1986, Os tman 2005), affecting population growth over t ime. Food limitation in insect populat ions has been suggested by severa l researchers based on laboratory exper iments (Jul iano 1986, Bommarco 1998), but it is difficult to extrapolate these results to populat ions in uncontrol led environments. Whi le food may be abundant in the environment, other factors such as energy expenditure and water st ress may prevent organisms from reaching their full reproductive,potential. Reproduct ion may a lso be affected by microcl imate changes , although not in such a direct manner. The main causa l factors for lowered reproduction in insects are related to obtaining and expending energy (Telfer and Butterfield 2004). To address how environmental condit ions affect a non-pest, forest-dwell ing spec ies , I se lected the most abundant spec ies of ground beetle in moist temperate forests of western North Amer i ca (Scaphinotus angusticollis) (Laval lee 1999, Lemieux and Lindgren 2004, R ichardson etal. in prep.) for study. Th is spec ies has been found by severa l studies to decl ine after forest harvesting (Leriski 1982, Laval lee 1999, Lemieux and Lindgren 2004); however s o m e management strategies may diminish the effects of logging on this spec ies (Chapter Three). Pas t studies have a lso found that it will persist in forest remnants after harvesting (Lemieux and Lindgren 2004), but there is 74 little ev idence to indicate what mechan isms may be responsib le for its pers istence or decl ine. By combining measures of feeding levels and energy storage, energy storage can be isolated a s a possib le causa l factor for di f ferences in reproductive status. Reproduct ion is only undertaken when energy avai lable from feeding is in e x c e s s of survival requirements (Perrin 1991). B e c a u s e there is a direct link from e x c e s s energy intake and reproductive capacity, reproductive status of individuals should be highest where feeding and fat storage coincide at high levels. Energy storage di f ferences between habitats may be accounted for by water st ress, particularly given its importance to insects (Todd and Block 1997). Whi le far more detai led measures of fat storage, hydration levels and reproductive status are avai lable in this study, compar isons of external body metrics can also be made with other studies that have examined live spec imens (e.g. Jul iano 1986, Bommarco 1998) and with data from previous years in this study (Chapter Two). In this study, a compar ison of the body condit ions of S . angusticollis is made between three levels of logging in riparian areas . B e c a u s e prior studies show that this beet le 's populat ions decl ine in c learcuts, I predict that all indices of condit ion will be higher with lowered amounts of logging. Mater ials and Methods Dissect ions In 2002, after the complet ion of the capture-recapture study (Chapter Three) , Scaphinotus angusticollis were captured using the s a m e trapping grids in the s a m e si tes, from May until November . From May through August , a control, 30m buffer and 75 clearcut site were sampled for one night every three weeks , with all nine sites visited over a nine-week period. In September , October and November , one control, 30m buffer and clearcut site were each sampled for one night per month. Beet les were placed in individual 7 ml sample vials and killed by freezing by placing in a -10 °C freezer within 2 hours of capture. Beet les were then continuously stored at -20 °C until February 2005, when d issect ions were begun. Al though some desiccat ion of spec imens is to be expected with this methodology, the relative water loss is a s s u m e d to be equal amongst spec imens . Dissect ions were made under a dissect ing microscope at 12x magnif icat ion. During d issect ions a number of different body condit ions were evaluated or ranked. Beet les were removed from the f reezer and placed on a pre-weighed a luminum dish. After reaching room temperature, e x c e s s condensat ion w a s removed by blotting with a paper towel and the beetles were weighed using an electronic counterbalance sca le (Sartorius B P 1 1 0 S ) to the nearest 0.1 mil l igram. After d issect ion, each beetle and tray was placed in a drying oven at 61 °C for 48 hours and then reweighed to obtain the dry weight. To determine the appropriate length of t ime to leave beetles to dry, seven sample beet les were dried in the oven and reweighed every other day for two weeks . In two of the spec imens tested, there w a s minimal loss in weight (-0.1 mg) after two days in the drying oven ; the others showed no weight loss after two days. Pronotum width and egg s i zes were measured during the d issect ion under the microscope using plastic dial cal ipers (accuracy 0.1 mm). Fat body, sex, sexual development, and foregut, midgut and hindgut fu l lness were all evaluated by ranking individuals into a numeric category. Fat body was ranked as 0 where none was evident; 76 this w a s only observed in teneral individuals. A ranking of 1 w a s ass igned where fat body occupied approximately 3 0 % of the body cavity and a ranking of 2 was ass igned where more than 3 0 % of the body cavity w a s taken up with fat body. Th is ranking system w a s deve loped from 12 preliminary d issect ions and represents the greatest degree of accuracy and precis ion in est imation. Sexua l development w a s ranked on the bas is of the sex organ development. Tenera ls were scored as 0; visible organs that did not have obvious accessory organ development or glandular function (visible as thick, white t issue) were scored as 1. Fully reproducing adults showed glandular t issues and sclerotization of supportive reproductive structures and they were ranked in category 2. Fema les had two addit ional categor ies in sexua l development. For gravid females, category 3 was used , in addit ion to counting and measur ing eggs found in the body cavity. Fema les with deve loped ovar ies that were crenated were p laced in category 4 and were considered to be expended females, having recently deposi ted their eggs. Paras i tes and any other factors (gut b lockages, internal damage) were noted as wel l . Anomal ies of gut shape and texture (e.g. gut hardenings at junctures between fore, mid and hind gut were noted in the d issect ions. The entire gastrointestinal tract was examined including c e c a e attached to the midgut. Gut contents of the foregut, midgut and hindgut were individually examined and descr ibed for texture and colour, to identify potential di f ferences in gut content appearances between habitat types. Compar i son to other studies of gut content examinat ions (Appendix I) were a lso possib le through these descr ipt ions. The three gut sect ions were ranked for "ful lness" as fol lows: "0" for no or very few (<10) particles and little or no fluid, " 1 " for larger numbers and s izes of particles or amounts of fluid, "2" for not iceable distention of the 77 gut sect ion, owing to fluid or particles contained within. Th is index w a s deve loped through observat ions made after 20 practice d issect ions were performed on individuals obtained from a feeding experiment for this thesis. In total, 193 beet les were d issected for this study. Al l spec imens trapped were d issec ted, with the except ion of only three samp les (denoted by aster isks (*) in Tab le 5.1). Owing to a large number of individuals caught for those three time periods, random subsampl ing of 12 to17 spec imens was done to reduce the number of d issect ions required. Weight / pronotum ratios Measu res of weight corrected for s ize have been used by many studies to evaluate the relative s ize of individuals (Jul iano 1986). Whether that mass is compr ised of fat, reproductive t issues, or gut contents, only d issect ions will reveal . Compar i son of a coarser measure of body condit ion (weight corrected for body s ize) can be made with a larger data set obtained in the 2000-2001 capture-recapture study outl ined in Chapter Two of this thesis. For this portion of the study, beet les were l ive-trapped in grids of pitfall traps set in forest, a clearcut and across a riparian buffer habitat of U B C ' s Ma lco lm Knapp Resea rch Forest. A more complete descript ion of trapping per iods and methods may be found in Chapter Two. Whi le individually marking beet les and noting their location on the grid of traps, beet les were also weighed with an electronic field sca le and their pronota measured using cal ipers. Pronotum width is used in this study as a proxy measure of s ize . The more common measure of insect s ize, head capsu le width, was not used because of the l ikelihood of damage to eyes , antennae and head scleri tes during measurement . T h e s e data were combined with the corresponding 78 measurements taken during the d issect ion of individuals caught from the s a m e sites in 2002. Al l data ana lyses were done using J M P (Version 4.04; Carey , N C ) . Resu l ts Life History Scaphinotus angusticollis adults were most active in May and September , and appeared to be able to overwinter as adults, as ev idenced by three recaptures of individuals from 2000 in the second field season . Others have suggested that they may also overwinter as eggs and larvae (Lindroth 1961). Sexua l d imorphism w a s evident in the d issected spec imens, with males the average weight w a s 0.17 g ( S E = 0.03) and average pronotum width was 0.36 c m ( S E = 0.02) (N = 121). Fema le average weight was 0.20 g ( S E = 0.06) and average pronota width w a s 0.37 c m ( S E = 0.02) (N = 128). Tenera l beet les were first detected in the population from mid-June through to September , with the peak in teneral emergence occurr ing in mid-July. Tenera ls were remarkable for their lack of a fat body, internal reproductive structures and support ive t issues. Y o u n g of the year ( Y O Y ) were easi ly identifiable even after sclerot izat ion of their integument, by brief examinat ion of the abdominal cavity, which conta ined no visible internal organs other than a digestive tract. G iven the means of trapping these individuals, tenerals must have some capacity for movement, despite their soft prothorax, the primary location of locomotor / musc le attachment. S o m e variability in emergence t imes between habitat types was noted and will be d i scussed below in further ana lyses . The 26 gravid females found in these d issect ions carried either one or two eggs , with either number of eggs approximately as common. The oblong, yel low eggs ranged 79 from 0.10 to 0.48 cm in length (mean = 0.42, S E = 0.22, N=37). Grav id females were only found in June and August to October, with the greatest range ac ross s e a s o n s in egg production occurring in the control habitat. Expended fema les were only found in August to October. Investment in reproductive structures (e.g. ovar ies, testes and accesso ry g lands) was substantial for both sexes and in full breeding condit ion, reproductive structures for both sexes compr ised approximately 6 0 % of abdominal cavity vo lume. A more complete d iscuss ion of reproductive development in beet les over the year and between habitats is presented below. Only one case of parasit ism was noted in all the d issect ions performed. A large parasit ic nematode measur ing approximately 80 mm in length w a s found in a young female taken from a clearcut site. Microparasi tes of the gut and Malpighian tubules could possibly have been present in the d issected individuals. Deformit ies and irregularities of the gut did not occur in more than 5 % of all individuals, precluding analys is for lack of data. Gut contents and fat body Average rankings of gut ful lness fol lowed very different patterns in the three habitats (Figure 5.1 A) . Beet les found in control si tes showed the lowest gut fu l lness and fat body ranking for this habitat in mid-season (July/August) (Figures 5.1 A and 5.1 B). Note that the average fat body ranking w a s three t imes higher in control si tes than in clearcuts at the beginning of the s e a s o n . Beet les from clearcuts started off the s e a s o n in May /June with their lowest gut ful lness and fat body rankings, with s teady increases over the year. In 30 m buffer habitats, beet les showed different trends for gut 80 fu l lness than for fat body, with steady inc reases in the average gut ful lness over the year, while the average fat body measure stayed remarkably constant at just over 1.0 (Figures 5.1 A and 5.1 B). By the end of the s e a s o n (September/October) , average values for both measures were very simi lar in all habitats. Prior to analys is , tests showed that these data were non-normal , even after transformation; tests therefore used rank data instead. Statist ical testing via A N O V A indicated that gut ful lness did not vary signif icantly with date (p = 0.78) (Table 5.2), but did general ly increase over the year (Figure 5.1 A) . Fat body w a s significantly inf luenced by habitat type (p = 0.03) (Table 5.3); a compar ison of the least square means suggests a significant difference (p = 0.02) between control and clearcut si tes (Figure 5.1 B). In genera l , fat body dec reased with more logging (Figure 5.1 B) . Multivariate analys is of var iance (dependent var iables gut fu l lness and fat body) indicated that there w a s a significant interaction of gut fu l lness and fat body with habitat (Wilks' Lambda p = 0.05) and date (Wilks' Lambda p = 0.02). W h e n compared to prior studies on gut contents of beet les (Appendix I), the quality of gut contents appeared to be very different from the plant material ingested in laboratory tests for feeding preferences. Throughout the s e a s o n s , gut materials were most consistent in quality and texture of feeding on soft-bodied prey like snai ls . Th is provides addit ional justification for the assessmen t of snai ls in these s a m e habitats (Chapter Four) and corroborates observat ions by other researchers (Lindroth 1961, Thie le 1977). Reproduct ive development 81 Beet les were ranked on sca les of 0-3 (males) or 0-4 (females), a s descr ibed in the Methods. To analyze the data appropriately, the number of categor ies had to be co l lapsed to a binary sys tem of either non-reproduct ive (1) or reproductive (2). Th is al lowed for males and females to be ana lyzed in the s a m e data set, which w a s necessary owing to the low number of individuals caught in some habitat types for some time periods. Reproduct ion in control and 30 m buffer habitats var ied slightly throughout the year, with large over lap between all treatments with no significant di f ferences in the least square m e a n s (Figure 5.2). Beet les in clearcut habitats showed an increase in reproduction throughout the year, fol lowing a distinctly different trend from the other habitats. Prior to statistical testing, data were arcs ine transformed, because binary data are inherantly proportional and require transformation. After transformation, data were found to be non-normal and thus statistical tests were run using rank data. Al though no significant effects were found with the main effects used in testing, the interaction of date and habitat showed the lowest probability of occurance to chance (p = 0.18) (Table 5.4); c loser scrutiny of the least square means for this interaction suggests that there is an earlier peak in reproductive status in 30 m reserves than in c learcuts and control si tes. Hydration levels Water content controlled for s ize (wet - dry weight/ wet weight) w a s general ly highest in May /June of 2002, but peaked in control s i tes in Ju ly /August (Figure 5.3). Water content of beet les w a s at a max imum in 30 m buffers in May /June at 3.1 g and at a minimum in c learcuts in September /October at 1.2 g. The least square means 82 indicate that clearcut si tes overal l had beet les with the highest water content, but no di f ferences were found to be significant (Figure 5.3). Variabil i ty of water content decl ined over the year in all habitat types. Prior to statistical testing, the data were found to be non-normal, even after log normal transformation, therefore A N O V A testing w a s run with ranked data. Stat ist ical testing detected a significant inf luence of some sites within certain habitats (p = 0.02) and a significant effect of date on water content (p = 0.04) (Table 5.5). Whi le the variation of beetle water content over time is evident (Figure 5.3), least square m e a n s for the nested term of site within habitat suggest that one site in e a c h of the three habitat types w a s very low and signif icanly inf luenced results. Weight/pronotum ratios In control s i tes, the ratio of we igh tpronotum width w a s highest in the early s e a s o n (May and June) , lowest in m idseason (July and August) , and increased again in the fall (Figure 5.4). A reverse trend w a s apparent in c learcuts, with peak weight:pronotum ratios occurr ing in Ju ly and Augus t (Figure 5.4). Al though distinct t rends occurred between habitat types, no significant relationship existed with habitat type (p = 0.21). Weight :pronotum ratios were significantly expla ined (p = <0.01) by the date of capture via an A N O V A . In subsequent testing via A N C O V A , In weight/pronotum ratios were found to significantly covary with fat body (p = <0.01) and water content (p = <0.01), but not gut fu l lness (p = 0.28). Reproduct ive rank was nearly significantly correlated (p = 0.06), with the interaction of sex development and habitat type significantly covarying with In weight pronotum ratio (Table 5.6). 83 Discuss ion Whi le s o m e anecdota l writings (Lindroth 1961, Thie le 1977) suggested that S . angusticollis may have low fecundity and be long-l ived, this study showed that previous est imates of 7 - 1 0 eggs per female (Thiele 1977) were optimistic, probably owing to prior studies taking place in laboratory condit ions with ample food avai lable. With only two breeding s e a s o n s readily apparent in the seasona l distribution of gravid females, and a max imum of two eggs produced per breeding s e a s o n , yearly production of eggs could not have exceeded four eggs per individual. E v e n with relatively high survival of adults and larvae, in order for any population of S. angusticollis to be sel f-replacing, adults would have to be long-l ived, surviving about four or five years. Th is suggests that only with very high survival of all life s tages, would populat ions of S . angusticollis self-sustain over t ime. Long- l ived adult s tages have been noted in other spec ies of large carabid beet les, using other measures of aging (Butterfield 1996). It is interesting to note that teneral emergence w a s relatively consistent, occurr ing in the early summer (July). Pupat ion and emergence of tenerals appears to be regulated by larger-scale factors (e.g. seasona l cl imate condit ions) that overwhelm any effect of the habitat di f ferences from logging. The most important compar isons in this study are of the internal condit ions of beet les, ac ross the habitat types examined. Water content varied more between sites than between habitats, suggest ing that individual site characterist ics may be more important to beetle water content than logging pract ices. Body water may have been inf luenced by other internal factors as wel l . Gut contents have a direct effect on the 84 amount of fluid present in the body of the beetle, as S. angusticollis tends to feed on water-rich prey items like snai ls (Lindroth 1961). T rends in gut fu l lness and fat body revealed two surprising aspec ts in the data. Whi le clearcut habitats and controls shared the s a m e trend in feeding and fat storage, 30 m buffers showed no correlation between the two body indices. Evidently clearcut si tes are general ly poorer than control sites in the early s e a s o n for both foraging and fat storage, but this trend is not as c lear later in the year. More difficult to explain are the results from the 30 m buffer si tes, where a different mechan ism appears to be dictating fat storage, relative to the amount of feeding. B a s e d on the amount of feeding, beet les in 30 m buffers appear to have greater fat stores than expected at the beginning of the s e a s o n and lower than expected fat stores at the end of the season . Other studies have found that fat body and feeding rates are directly and positively correlated (Jul iano and Lawton 1990) If there is a disjunct between the rate of food consumpt ion (gut ful lness) and fat storage, this may have implications for reproduction, which is a lso correlated with fat body in most beet les (Bommarco 1998). With a higher amount of fat body in 30 m buffer si tes, this could result in an increase in fecundity and a resulting increase in population s ize in this type of habitat. Whi le the fecundity in different habitats could not be examined in this study, the timing of reproduction appears to be slightly, but not significantly different in the three habitats. T h e s e results should be interpreted carefully, however, a s there may be individuals caught who have come into a habitat from other areas , (i.e., their condit ion may not be dictated by the microcl imate, prey, or competitors in the site they are found). Large amounts of movement were noted in these 30 m 85 buffers in 2000 and 2001 (Chapter Two), so it is possib le that the effect of the habitat on the internal condit ions of individuals was confounded by non-resident beetles. Internal condit ions of the beet les in this study indicate that while clearcut and 30 m buffers may harbour similar numbers of beet les (Chapter Two), internal condit ions of beet les in these habitats may not be equal . This finding underscores the importance of studying populat ions at a greater level of detail than p resence /absence studies have to offer. S t ressors such as food shortages and changing microcl imate will manifest themse lves first in the physical condit ion of individuals (Perrin 1991) and may later become detectable as inc reases in death or emigration rates. F e w insect eco logy studies have examined internal condit ions of insects (e.g. fat body, gut ful lness) and how they fluctuate in different habitats, with s o m e notable except ions (Jul iano 1986, Bommarco 1998, Os tman 2005). Internal condit ion may have short-term consequences on individuals such as their ability to survive adverse condit ions, evade predators, or resist parasit ic infection, but long-term consequences such as lower quality (size) and number of offspring (Heg and R a s a 2004). Compar i son of the internal measures of body condit ion (reproduction, water content, gut ful lness, and fat body) with the more readily avai lable external measures (weight/pronotum ratios) indicate that some internal measures (fat body and water content) are tightly correlated with weight/pronotum ratios. With respect to the practicalit ies of field work and the minimization of the impacts of studies, external body metrics are vital and useful information to gather. F rom a theoretical perspect ive however, ascr ibing weight/pronotum ratios to any particular body condit ion is not poss ib le without further ev idence, making it hard to draw conc lus ions about internal 86 condit ions from external measures . Ideally, studies that use body condit ion to compare habitats should use several internal measures that suggest different mechan isms for population decl ines. For example , concurrent measurement of gut fu l lness and fat body indicated a very different set of condit ions in 30 m buffers in this study. Th is study provides a vital link to an important area of insect eco logy that de lves beyond p resence /absence and into the manifestation of physiological responses of o rgan isms to the environment. Th is study represents the first to use internal condit ions of an insect as a measure of habitat suitability in forests. A s an indicator of disturbance, my results suggest that while providing s o m e advantages over clearcut habitats, 30 m buffers are fundamental ly different from cont iguous forest. 87 Literature Ci ted Bennett, V .A . , O. Kuka l , and R . E . Lee . 1999. Metabol ic opportunists: Feed ing and temperature influence the rate and pattern of respiration in the high arctic wool lybear caterpil lar Gynaephora gornlandica (Lymantri idae). Journal of Exper imental Biology 202 : 47-53. B lanckenhorn, W . U . 2006. Divergent juvenile growth and development mediated by food limitation and foraging in the water strider Aquarius remigis (Heteroptera: Gerr idae). Journal of Zoo logy 268:17-23. Boggs , C . L . and K.D. F reeman . 2005. Larval food limitation in butterflies: effects on adult resource al location and f i tness. Oeco log ia 144: 353-361. Bommarco , R. 1998. S tage sensitivity to food limitation for a general ist arthropod predator, Pterostichus cupreus (Coleoptera: Carab idae) . Environmental Entomology 27: 863-869. Butterfield, J . 1996. Carab id l ife-cycle strategies and cl imate change: A study on an altitude transect. Eco log ica l Entomology 21 : 9-16. C h e n , J . , S . C . Saunders , T .R . Crow, R . J . Na iman, K.D. Brosofske, G . D . Mroz , B.L. Brookshire, and J . F . Frankl in. 1999. Microcl imate in forest ecosys tem and landscape ecology. B iosc ience 49: 288-297. Dmitriew, C . and L. Rowe. 2005 . Resource limitation, predation risk and compensatory growth in a damself ly. Oeco log ia 142: 150-154. Heg , D. and O . A . E . R a s a . 2004. Effects of parental body condit ion and s ize on reproductive s u c c e s s in a tenebrionid beetle with biparental care. Ecolog ica l Entomology 29: 410-419. 88 Jul iano, S . A . 1986. Food limitation of reproduction and survival for populat ions of Brachinus (Coleoptera: Carab idae) . Eco logy 67: 1036-1045. Ju l iano, S .A . and J . H . Lawton. 1990. Extr insic vs . intrinsic food shortage and the strength of feeding links: effects of density and food availability on feeding rate of Hyphydrus ovatus. Oeco log ia 83: 535-540. Lava l lee, S .L . 1999. C h a n g e s in the carabid beetle community (O. Co leoptera , F. Carab idae) of the S i c a m o u s Creek R e s e a r c h Site. Master 's thesis, Dept. of Zoology, U B C . P p . i to 73. Lemieux, J . P . and B . S . L indgren. 2004. Ground beetle responses to patch retention harvesting in high elevat ion forests of British Co lumbia . Ecography 27: 557-565. Lensk i , R . E . 1982. The impact of forest cutting on the diversity of ground beetles (Coleoptera: Carab idae) in the southern Appa lach ians . Eco log ica l Entomology 7: 385-390. Lindroth, C . H . 1961. The Ground Beet les of C a n a d a and A l a s k a . Opuscu la Entomologica Supp lementum X X - X X I V . P p . 1-1192. Lovejoy, D.A. and S . J a h a n . 2006. Phytogeny of the cort icotrophin-releasing factor family of peptides in the metazoa. Genera l and Compari t ive Endocr inology 146: 1-8. Moehr l in, G . S . a n d S . A . Ju l iano. 1998. Plasticity of insect reproduction: testing models of flexible and fixed development in response to different growth rates. Oeco log ia 115 :492 -500 . 89 Ostman , O. , B. Ekbom, J . Geng tsson , and A . - C . Weibu l l . 2001 . Landscape complexi ty and farming practice influence the condit ion of po lyphagous carabid beet les. Eco log ica l Appl icat ions 11: 480-488. Os tman , O. 2005. Asynchronous temporal variation among sites in condit ion of two carabid spec ies . Ecologica l Entomology 30: 63-69. Perr in, N. 1991. Opt imal resource al location and the marginal value of organs. T h e Amer ican Naturalist 139: 1344-1369. Rick lefs, R . E . and M. Wike lsk i . 2002. The physiology/l i fe history nexus. T rends in Eco logy and Evolut ion 17: 462-468. Romero , L . M , 2004. Physio logical stress in ecology: lessons from biomedical research. Trends in Eco logy and Evolut ion 19: 249-255. S p e n c e , J . R . and N iemela , J .K . 1994. Sampl ing carabid assemb lages with pitfall traps the madness and the method. The Canad ian Entomologist 126: 881-894. Statzner, B., A . G . Hildrew, and V . H . R e s h . 2001. S p e c i e s traits and environmental constraints: entomological research and the history of ecological theory. Annua l Rev iew of Entomology 46: 291-316. Telfer, G . and J . E . L . Butterfield. 2004. The control of reproductive d iapause in Nebr ia sal ina (Coleoptera: Carab idae) . Ecolog ica l Entomology 29: 482-487. Thie le, H.U. 1977. Carab id Beet les in their Environments. Spr inger Ver lag , Ber l in. Pp . xvi i-369. Todd , C M . and W . Block. 1997. R e s p o n s e s to desiccat ion in four coleopterans from sub-Antarct ic South Georg ia . Journal of Insect Phys io logy 43 : 905-913. 90 Wikelsk i , M. and S . J . C o o k e . 2006. Conservat ion physiology. T rends in Eco logy and Evolut ion 21 : 38-46. 91 Tab les Table 5.1 Summary of trapping dates and sites, with total number of Scaphinotus angusticollis dissected per time period and site. Zeroes indicate that no individuals were caught. Date Site Name Treatment Number caught/ Number dissected May 31 Control 2/2 30 m reserve 3/3 Clearcut 0/0 June 6 Control 4/4 30 m reserve 4/4 Clearcut 2/2 June 12, 2002 Moss Control 16/16 South Creek 30 m reserve 23/23 B Clearcut 2/2 June 25, 2002 Control 1/1 30 m reserve 0/0 Clearcut 8/8 July 11, 2002 Moss Control 5/5 South Creek 30 m reserve 27/17* B Clearcut 3/3 July 18, 2002 Mike Control 1/1 H 30 m reserve 1/1 E Clearcut 2/2 August 8, 2002 Mike Control 1/1 South 30 m reserve 0/0 B Clearcut 0/0 August 15, 2002 G Control 0/0 D 30 m reserve 11/11 I Clearcut 0/0 August 22, 2002 Moss Control 8/8 South 30 m reserve 26/26 B Clearcut 0/0 September 19, 2002 G Control 16/16 D 30 m reserve 44/13* I Clearcut 9/9 October 21, 2002 Mike Control 27/27 South 30 m reserve 23/23 E Clearcut 56/12* *denotes site/dates for which all individuals collected were not all analyzed. 92 Table 5.2 Analysis of variance of gut fullness for Scaphinotus angusticollis in forest, 30 m reserve, and clearcut sites. Main effects include habitat type (H), site nested within habitat, date and the interaction of date and habitat type. (Model r2 = 0.38, df = 14) Source DF Sum of Squares MS F-Ratio P Habitat (H) 2 57.53 28.77 1.81 0.24 Site (H) 6 95.62 15.94 0.34 0.90 Date 2 23.86 11.93 0.26 0.78 Date*H 4 164.08 41.02 0.89 0.50 Error 12 555.40 46.28 Total 26 896.50 Table 5.3 Analysis of variance of fat body for Scaphinotus angusticollis in forest, 30 m reserve, and clearcut sites. Main effects include habitat type (H), date nested within habitat type, and site nested within habitat type. (Model r2 = 0.45, df = 14) Source df Sum of Squares MS F-Ratio P Habitat (H) 2 72.80 36.40 7.03 0.03 Site (H) 6 31.05 5.17 0.27 0.94 Date 2 29.67 14.83 0.77 0.47 Date*H 4 57.97 14.49 0.75 0.58 Error 12 232.09 19.34 Total 26 423.57 Table 5.4 Ordinal logistic fit for rank of sex development for Scaphinotus angusticollis in forest, 30 m reserve and df Sum of Squares MS F-Ratio P 93 clearcut sites. Main effects include habitat type (H), date nested within habitat type, and site nested within habitat type. (Model r 2 = 0.43, df= 14)Source • Habitat 2 8.35 4.17 1.45 0.31 Site (H) 6 17.33 2.89 0.11 0.99 Date 2 14.70 7.35 0.28 0.79 Date * H 4 193.15 48.29 1.86 0.18 Error 12 311.85 25.99 Total 26 545.37 Table 5.5 Analysis of variance results of water content of Scaphinotus angusticollis in forest, 30 m reserve and clearcut sites. Main effects include habitat type (H), site nested within habitat, date and the interaction of date and habitat. (Model r2 = 0.74, df = 14) Source df Sum of Squares MS F-Ratio P Habitat (H) 2 9.63 4.81 0.04 0.96 Site (H) 6 646.26 107.71 3.90 0.02 Date 2 243.88 121.94 4.41 0.04 Date * H 4 21.12 5.28 0.19 0.94 Error 12 331.45 27.62 Total 26 1252.33 94 Table 5.6 Results for analyses of covariance tests of weight/pronotum ratio with gut fullness, fat body, water content and sex Scaphinotus angusticollis. Sou rce D F M S F Va lue P r > F S e x Development (S) 1 0.009 4.23 0.06 Habitat Type (H) 2 0.01 5.45 0.02 S * H 2 0.01 5.11 0.02 Gut fu l lness (G) 1 0.004 1.27 0.28 Habitat Type (H) 2 0.005 1.42 0.28 G * H 2 0.004 1.3 0.31 Fat Body (F) 1 0.03 25.51 0.0002 Habitat Type (H) 2 0.002 1.19 0.34 F*H 2 0.002 1.4 0.28 Water content (W) 1 0.03 25.35 0.0002 Habitat (H) 2 0.0004 0.28 0.76 W * H 2 0.0004 0.27 0.77 95 Figures 5.0 4.5 H 8 4.0 H CD *! 3.5 O) "S 3.0 -4—' c CO © • 2.5 2.0 1 • Control V 30 m reserve • Clearcut May/June V? July/August Date September/ October Least square means Figure 5.1 Mean total gut fullness (A) and fat body (B) in beetles from control, 30 m buffer and clearcut sites in 2002. N= 2-3, depending on data point; error bars indicate standard error. Asterisk (*) indicates significant difference. 96 c • CO CD > o "O O i _ CL CD c 03 0 2.4 2.2 2.0 1.8 1.6 1.4 1.2 1.0 0.8 # Control ^ 30 m buffer • Clearcut • • May/June July/August Date September/ Least October square means Figure 5.2 Mean binary reproductive rank of beetles in control, 30m buffer and clearcut habitats. N = 2-3, depending on data point; error bars indicate standard error. 3.5 3.0 A 2.5 3 c CD O 2.0 CD 03 1.5 1.0 0.5 0 • Control 30 m buffer • Clearcut 0 *7 a May/June July/August September/ Least October square Date means Figure 5.3 Water content of beetles controlled for body size (wet — dry weight/ wet weight) from control, 30 m buffer and clearcut habitats in 2002. N = 2-3, depending on data point; error bars indicate standard error. 97 0.65 H 0.60 A 0.55 0.50 .2> 0.45 — 0.40 0.35 H u • • Control 30 m buffer Clearcut — i — 10 11 Month Figure 5.4 M e a n ln(weight+1 )/pronotum ratios (g/cm) for beet les caught in control, 30 m buffer and clearcut habitats in 2000-2002 . N = 2-3, depend ing on data point; error bars indicate s tandard error. Months cor respond directly to the ca lendar months (i.e. 5=May, 6=June, etc.). 98 Chapter Six Conclusion Findings of this study Ripar ian reserves may have benefits for terrestrial organisms that prefer to live in forested habitat. Ripar ian management has been focused mainly on vertebrates (Richardson et al. 2005), but treed reserves provide adequate habitat for some forest-preferring, non-pest insects such as Scaphinotus angusticollis. However, d is tances moved by S. angusticollis suggest that individuals in buffers perceive or use this buffer habitat differently than cont iguous forest. Whi le no direct interpretation of habitat suitability can be made where individuals have moved greater d is tances, this is indicative of dif ferences in habitat percept ion (Szyszko et al. 2004). Th is cal ls to quest ion whether number of individuals a lone is adequate information to a s s e s s habitat suitability for insect populations. Food resources are one of the most important factors limiting predacious insect populat ions (Jul iano 1986, Os tman 2005) and may have an effect on individual behaviour as wel l . Surveys for snai ls were conducted because they typically are considered to be S . angusticollis' main food source. T h e s e surveys showed that while abundances are similar between forested, buffer and clearcut habitat, subtle shifts in spec ies occur after harvesting. Whi le feeding studies were unsuccessfu l in determining a preference for snai l spec ies , other potential food sources present in forest habitat (e.g. severa l spec ies of Rubus) were identified. Food resources are present in all three habitat types examined in this study, which cal ls to quest ion whether individuals exploit resources differently in these habitats, consuming resources proportionally to their 99 availability. If resources are of unequal nutritional value to S. angusticollis, this may result in lower population s i zes in some habitats. Internal body condit ions provide valuable information on how individuals consume, store and convert food resources (Ricklefs and Wikelsk i 2002). T h e s e condit ions a lso provide insight into potential mechan isms for population decl ine, as lower fat storage, fecundity and changed reproductive schedu les may result from food shortages or shifts in the temporal availability of food resources. Internal condit ions of S . angusticollis in control, reserves and clearcut si tes suggest that fat storage and gut ful lness (proxy measure for feeding) are not tightly correlated in reserve sites. Th is may have resulted from higher energy expenditure in this habitat or migration into this habitat by individuals from the surrounding clearcut. Th is study illustrates the utility of body condit ion as a measure of habitat suitability for insects. It is surprising that more studies in the field of appl ied insect ecology have not employed these wel l-known methods, particularly when they may be able to indicate mechan isms for changes in populat ions. Instead, most non-pest invertebrate research has examined how community shifts take place after habitat change. F rom the results of this study, measures of weight corrected for s ize are good indicators of some internal condit ions (e.g. fat body and water content), but not others (reproductive status and gut ful lness). Caut ion should be used when extending connect ions between body condit ions to other spec ies , though. Stud ies that examined severa l spec ies of carabid beet les for their responses to environmental condit ions found that there are vast di f ferences in fat body and reproductive investments between spec ies (Ostman et al. 2001). Pilot data should be col lected to verify what internal 100 measures are best est imated by external condit ions for each spec ies to be studied. Then a simplif ied field methodology using external measures can be used to est imate habitat suitability for a spec ies . U s e of riparian reserves by forest-preferring spec ies In the harvested landscape, riparian reserves may compr ise a large proportion of the unharvested forest landscape. There have been few studies to quantify the use of riparian reserves by terrestrial invertebrates, despite the importance of insects to the forested ecosys tem (Taylor and Doran 2001) and their potential as indicators of habitat d isturbance (Koivula and N iemela 2003) and its impacts on the forest biota. In Brit ish Co lumb ia , riparian reserves are not only intended to isolate s t reams from upslope land uses , but they are a lso cons idered to provide habitat for forest dwel lers (Anonymous 1995), despite the paucity of empir ical ev idence for this extension (Hylander et al. 2004). Th is study illustrates the potential for riparian reserves to be used a s habitat by forest-preferring spec ies . B e c a u s e some terrestrial insects such as S . angusticollis are f l ightless, with low d ispersal capabil i t ies, their needs may differ greatly from vertebrate counterparts (Hylander et al. 2004), although compared to other insects, carabids may have relatively fast recoIonization rates (Brouat et al. 2004). Al though this study illustrates some benefit of buffers to this spec ies ' abundance , it should be noted that this study was conducted on headwater s t reams of relatively smal l s i ze . Un less headwater s t reams bear f ish, they would not normally receive protection from logging under the current legislation in Brit ish Co lumb ia (Moore 2005). The taxa and habitat of study represent a specia l ized c a s e , for both political and biological reasons. G iven the low priority afforded to insect conservat ion (Samways 101 2006), it is doubtful that advocacy for riparian forest remnants as terrestrial insect habitat, would ever reach fruition. E v e n amongst conservat ionists, the political cl imate for arthropod conservat ion is poor (Kellert 1993). In addit ion, the riparian habitat of headwater s t reams in the Paci f ic Northwest differs from that of higher order s t reams and in different regions (Richardson et al. 2005), as there is little transition from riparian community to upslope forest habitat. Th is suggests that forest-preferring spec ies such as S. angusticollis may not perceive the presence of smal ler s t reams when in cont iguous forest. This assert ion is supported by the f indings of this study, which found no significant di f ferences in population s i zes or movements of S. angusticollis in 30 m buffers. U s e of indicator taxa Indicator taxa are of value to broad-scale survey studies that seek to determine landscape- level patterns and compare between wide-ranging habitats (Caro and O'Doherty 1999). Within most entomological studies, surveys unfortunately involve little more information than p resence /absence data and relative densi t ies. T h e results of this study highlight the importance of more detailed information on movement within populat ions, and suggest that some internal measures of condit ion (e.g. fat body and gut ful lness) may be eas ier to interpret than others (e.g. relative water content) using external measures . Whi le some studies have used internal factors such as these to evaluate agricultural p rocesses (Ostman et al. 2001), there is an opportunity for greater appl icat ions to forestry as well . There may be very different internal condit ions between spec ies in the s a m e environmental condit ions (Ostman 2005), suggest ing that broader conc lus ions regarding habitat suitability for other spec ies may not be determined from 102 just one spec ies . Ideally, a compar ison of internal condit ions in severa l forest-dwell ing spec ies would provide greater insight into how widespread the applicabil i ty of this spec ies ' internal condit ions are as an indicator of habitat suitability. Contr ibutions to insect ecology and conservat ion S p e c i e s such as Scaphinotus angusticollis may make good indicators of habitat d isturbance, as they are w idespread, easy to identify and their populat ions readily reflect habitat changes like logging. Scaphinotus angusticollis provides a different perspect ive on the terrestrial insect community, represent ing longer- l ived, low fecundity insects. Whi le this study may represent an extreme in the typical life history of insects, it is important to cons ider the range of eco log ies and life histories when examining the impacts of habitat d isturbance on the biota (Hylander et al. 2004). A l though there have been some studies on the eco logy of larger, longer-l ived beet les (Butterfield 1996), few studies of insect diversity acknowledge that spec ies responses may differ widely on the basis of their life history. Conservat ion planning for longer-l ived spec ies is receiving greater attention because of the specia l problems involved in monitoring spec ies with longer life spans . Populat ions may continue to decl ine after the s t ressor is removed (Noon and B lakes ley 2006), short-term population growth fates may be more difficult to est imate (Koons et al. 2005), and life history character ist ics may need to be wel l -known before larger-scale patterns can be interpreted (Kalwij etal. 2005). A l though insects are commonly assumed to have intermediate response t imes to plants and micro-organisms (Hodkinson and J a c k s o n 2005), variability in life histories may result in differing approaches for conservat ion planning. 103 Populat ions of S . angusticollis may ass is t in defining the minimum sca le at which planning should incorporate the needs of non-pest, forest-dwell ing insects. In terms of risk, large and fl ightless insects are more likely to become extirpated (Brouat et al. 2004) and may be eas ier to conserve, owing to their smal ler range (P imm and R a v e n 2000). P lanning that incorporates invertebrates may help at the smal ler sca le of planning conservat ion a reas that Harris et al. (2005) advocate , which requires more detai led analys is of land cover and current land uses . Indications from this study suggest that for ground-dwell ing insects like S . angusticollis, populat ions in 30 m riparian buffers are different in quality but not quantity. F rom this thesis, I argue for a change in our approach to insect monitoring and study. Far more information regarding habitat suitability can evidently be gained by looking at the internal and external condit ions of individuals, with minimal effort. Al though this study incorporated severa l measures of internal and external body condit ion, a modified field procedure would reduce the amount of effort required. A s opposed to relative abundances , addit ional information on reproductive condit ion and levels of feeding dist inguish high densi t ies of individuals from thriving populat ions. Quest ions such as those being posed for vertebrates in conservat ion physiology (Wikelski and C o o k e 2006) could be addressed for insect conservat ion, where low numbers of individuals may preclude sampl ing or renders it inadvisable. In addit ion, we should be call ing to quest ion the w idespread sampl ing and killing of potentially long-lived insect spec ies . Many studies have documented the negative effects of research, from the spread of d i sease of handl ing (Ginsberg et al. 1995), to decl ining survival through radio-collaring (Moorhouse and Macdona ld 2005) and 104 introduction of fungi and d i sease through tree core sampl ing (Phil l ips et al. 1998). Even methods such as catch and re lease have been assoc ia ted with subsequent dec l ines in survival in butterflies (Mallet e r a / . 1987). Whi le such methods certainly have lower impacts on their target spec ies than catch-kil l methods, planning of sampl ing and monitoring needs to cons ider the impact on target spec ies . 105 Literature Ci ted Anonymous . 1995. Ripar ian Management A rea Guidebook. Brit ish Co lumb ia Ministry of Envi ronment and British Co lumb ia Ministry of Forests . Forest Pract ices C o d e of Brit ish Co lumb ia . Brouat, C , S . Meusnier , and J . - Y . Rasp lus . 2004. Impact of forest management pract ices on carabids in European fir forests. Forestry 22 : 85-97. Butterfield, J . 1996. Carab id l ife-cycle strategies and climate change: A study on an altitude transect. Eco log ica l Entomology 21 : 9-16. Ca ro , T . M . and G . O'Doherty. 1999. O n the use of surrogate spec ies in conservat ion biology. Conservat ion Bio logy 13: 805-814. G insberg , J . R . , K.A. A lexander , S . C ree l , P .W. Kat, J . W . McNutt, and M.G.L . Mil ls. 1995. Handl ing and survivorship of Afr ican wild dog {Lycaon pictus) in five ecosys tems . Conservat ion Biology 9: 665-674. Harris, G . M . , C . N . Jenk ins , and S . L . P i m m . 2005. S c a l e of planning conservat ion areas . Conservat ion Biology 19: 1957-1968. Hodk inson, I.D. and J . K . J a c k s o n . 2005. Terrestrial and aquatic invertebrates as bioindicators for environmental monitoring, with particular reference to mountain ecosys tems . Environmental Management 35 : 649-666. Hylander, K., C . N i lsson, and T. Gothner. 2004. Effects of buffer-strip retention and clearcutt ing on land snai ls in boreal riparian forests. Conservat ion Biology 18: 1052-1062. Ju l iano, S . A . 1986. Food limitation of reproduction and survival for populat ions of Brachinus (Coleoptera: Carab idae) . Eco logy 67: 1036-1045. 106 Kalwij , J . M . , H.H. Wagner , and C . Sche idegger . 2005. Effects of stand-level d is turbances on the spatial distribution of a l ichen indicator. Ecolog ica l Appl icat ions 15: 2015-2024. Kellert, S . R . 1993. Va lues and percept ions of invertebrates. Conservat ion Biology 7: 845-855. Ko ivu la , M. and J . N iemela . 2003. G a p felling as a forest harvesting method in boreal forests: responses of carabid beet les (Coleoptera, Carab idae) . Ecography 26: 179-187. Koons , D.N. , J . B . G rand , B. Zinner, and R .F . Rockwel l . 2005 . Transient population dynamics : Relat ions to life history and initial population s ize . Ecolog ica l Model l ing 185 :283 -297 . Mallet, J . , J .T. Longino, D. Murawski , A . Murawski , and A . S i m p s o n De G a m b o a . 1987. Handl ing effects in Heliconius: Where do all the butterflies go? Journal of An imal Eco logy 56: 377-386. Moore , R.D. 2005. Smal l s t ream channels and their riparian z o n e s in forested catchments of the Paci f ic Northwest: Introduction. Journal of the Amer ican Wate r Resou rces Assoc ia t ion 41 : 759-761. Moorhouse , T . P . and D.W. Macdona ld . 2005. Indirect negative impacts of radio-col laring: sex ratio variation in water vo les. Journal of Appl ied Eco logy 42: 91-98. Noon , B .R. and J .A . B lakes ley . 2006. Conservat ion of the Northern Spotted Owl under the Northwest Forest P lan . Conservat ion Biology 20: 288-296. Os tman , O. 2005. Asynchronous temporal variation among sites in condit ion of two carabid spec ies . Ecolog ica l Entomology 30: 63-69. 107 Ostman , O., B. Ekbom, J . Geng tsson , and A . - C . Weibu l l . 2001 . Landscape complexity and farming practice inf luence the condition of po lyphagous carabid beet les. Eco log ica l Appl icat ions 11: 480-488. Phi l l ips, O.L. , P. Nunez V . , and M . E . T imana . 1998. Tree mortality and col lect ing botanical vouchers in tropical forests. Biotropica 30: 298-305. P i m m , S .L . and P. R a v e n . 2000. Extinction by numbers. Nature 403 : 843-845. R ichardson, J . S . , R . J . Na iman, F . J . S w a n s o n and D .E . Hibbs. 2005. Ripar ian communit ies assoc ia ted with Paci f ic Northwest headwater s t reams: assemb lages , p rocesses , and un iqueness. Journal of the Amer ican Water Resou rces Assoc ia t ion 41 : 935-947. Ricklefs, R . E . and M. Wike lsk i . 2002. The physiology/l i fe history nexus. T rends in Eco logy and Evolut ion 17: 462-468. S a m w a y s , M .J . 2006. Insect extinctions and insect survival. Conservat ion Bio logy 20: 245-246. S z y s z k o , J . , S . Gryuntal , and A . Schwerk . 2004. Dif ferences in locomotory activity between male and female Carabus hortensis (Coleoptera: Carab idae) in a pine forest and a beech forest in relation to feeding state. Environmental Entomology 33: 1442-1446. Taylor, R . J . and N. Doran. 2001 . U s e of terrestrial invertebrates as indicators of the ecological sustainabil ity of forest management under the Montreal P r o c e s s . Journal of Insect Conservat ion 5: 221-231. Wikelsk i , M. and S . J . C o o k e . 2006. Conservat ion physiology. Trends in Eco logy and Evolut ion 21 : 38-46. 108 Appendix I Food preference in Scaphinotus angusticollis (Fischer Von Waldheim) Introduction For this study, food preferences of the carabid beetle Scaphinotus angusticollis (F ischer V o n Waldhe im) were examined using prey that are avai lable in forests and clearcuts of Paci f ic Northwestern coniferous forests. Th is spec ies w a s chosen because of prior studies on its feeding ecology that hold S. angusticollis as a prime example of a special ist predator, consuming primarily gastropods, pulmonate snai ls in particular (Lindroth 1961, Thiele 1977, Digweed 1993). Whi le there has been little empir ical examinat ion of the strength of the trophic relationship between S. angusticollis and pulmonate snai ls, a congener (S . marginatus) has been shown to be highly select ive in food cho ices , with evident preference for certain spec ies (Digweed 1993). Speci f ical ly, this feeding behaviour study was des igned to evaluate potential trophic relat ionships between S. angusticollis and severa l potential food sources present in the natural environment. The initial hypothesis posed in this study was : Do S. angusticollis significantly prefer gastropod prey over other food cho ices avai lable in their envi ronment? For the second experiment, hypothesis tested was : Do S. angusticollis select to consume other forms of plant matter found in their habitat? Methods Bas i c experimental des ign Scaphinotus angusticollis beet les were l ive-trapped using standard pitfall traps on the night of June 24, 2002 (Experiment One) and July 8, 2002 (Experiment Two) in two unharvested experimental si tes in Malco lm Knapp R e s e a r c h Forest. Beet les were 109 brought back to the laboratory at U B C and p laced into separate containers. Fifteen Rubbermaid ® 10 litre tote boxes with lids provided individual a reas for beet les to live for the duration of the experiment. In each container, approximately 3 litres of moist, sterile potting soil lined the bottom, providing a semi-natural substrate. A 10 c m x 10 c m piece of step moss (Hylocomium splendens) w a s placed on the potting soi l , to provide a hiding spot for beet les. Conta iners were misted daily for the duration of the experiment, and light condit ions were set to the current daylight hours using an automated timer. Temperature condit ions in the lab ranged from 18 to 21° C . In each container, beet les were concurrently provided with a variety of food cho ices in a "cafeteria" style food choice experiment (Chesson 1978). In this type of experimental des ign , the first food consumed in part or whole is cons idered to be the first "choice". At this point, the experiment may be terminated, but in the case of this experiment, beet les were kept in containers and any further responses to food were monitored. Beet les were weighed every day, using an electronic scale (Ohaus Scout , cal ibrated to 1.00 g), to provide some confirmation of food consumpt ion. At the end of the experiment, all beet les were p laced into individual vials and killed by freezing. Dissect ion of the beet les used in the experiment was further used to confirm food consumpt ion during the course of the experiment. Food cho ices Select ion of potential food items was made based on publ ished studies of Scaphinotus angusticollis and its congener 's food preferences. These studies (Lindroth 1961 and Thiele 1977) lacked empir ical data, with one except ion (Digweed 1993). For 110 the first experiment, beet les were provided with one isopod (Armadillidium vulgare), one medium-s ized pulmonate snai l (either Ancotrema sp . or Vespericola columbianus). Lettuce was a lso placed in the container, to sustain the isopod and the snai l until the experiment was terminated. For the second experiment, beet les were first provided with a snai l , a sow bug and some lettuce for five days, and then new plant matter was provided for consumpt ion. Beet les received fruit from raspberr ies (Rubus sp) or blueberr ies (Vaccinium sp.) obtained from Malco lm Knapp Resea rch Forest or lettuce purchased from the grocery store. The three cho ices were selected to represent varying degrees of novelty to the beetles. Whi le the first experiment was intended to evaluate beetle food choice between live, animal prey, the second experiment was des igned to falsify f indings from the first feeding experiment using different plant foods. Data analys is No statistical analys is of these results is presented here. Resul ts Resu l ts from the initial test of food preference by S . angusticollis showed that every beetle subjected to testing selected lettuce over all other prey i tems present in containers. Th is result was not expected, as the lettuce w a s placed in the container to sustain the animal prey items. Feed ing on the lettuce w a s noted on the first day after placing beet les in containers, and continued for all five days of the experiment. To corroborate these unexpected f indings, all spec imens were subsequent ly killed by freezing at the end of the experiment and d issected for examinat ion of gut contents. In 111 every c a s e , the mid and hindgut of each beetle contained green plant material that w a s easi ly identified as lettuce particles, as no other food items of this colour and cons is tency were present. Statist ical ana lyses of these data would be superf luous, a s the demonstrat ion of preference for lettuce is obv ious and pronounced in this experiment. In the second experiment, only one beetle chose to eat lettuce, whereas four beet les exposed to raspberr ies fed readily on them and all five beet les exposed to blueberr ies fed as well (Figure One) . D iscuss ion Scaphinotus angusticollis appears to have more catholic feeding habits than has been previously reported in the literature. Plant matter, in the form of lettuce, w a s readily consumed by beet les in the initial experiment, and berries were a lso consumed in the second experiment performed. Whi le these f indings may initially s e e m to completely contradict prior studies on S. angusticollis, this study may be expla ined by other factors. O n e possib le explanat ion for this unexpected response may lie in the life history of S . angusticollis and cl imatic condit ions they were gathered under. In early July, when these beet les were col lected, many individuals are preparing for the hot and dry months of July and August , when this particular study site receives little or no rainfall (Kiffney et al. 2003). During this t ime, S. angusticollis probably go into aest ivat ion, reemerging in the fall activity period (Lindroth 1961, Thie le 1977). It is not surprising that beet les caught during this time period might choose from avai lable foods on the bas is of water content, as is consistent with these results. 112 These f indings in no way discredit the justification for studying snai l diversity in the riparian habitat types examined in this thesis (Chapter Four). A s w a s observed in the d issect ion of individuals in Chapter F ive, very different materials are present in the guts of wild-caught beet les. T h e s e materials were descr ibed a s being most consistent in texture and colour of animal material, such as snai ls or s lugs (Chapter Five). The overal l implication of this experiment was that there are many potential food resources that S . angusticollis will exploit in its environment; some have not yet been descr ibed in the publ ished literature. It is possib le that at differing t imes of the year, S . angusticollis relies on food resources that are higher in water or protein, a s hot cl imate or egg production may require addit ional nutrients. 113 Literature Ci ted C h e s s o n , J . 1978. Measur ing preference in select ive predation. Eco logy 59: 211-215. D igweed, S . C . 1993. Select ion of terrestrial gastropod prey by Cychr ine and Pterost ichine ground beet les (Coleoptera: Carab idae) . The C a n a d i a n Entomologist 125: 463-472. Kiffney, P . M . , J . S . R ichardson, and J . P . Bul l . 2003 . R e s p o n s e s of periphyton and insects to manipulation of riparian buffer width a long forest s t reams. Journa l of Appl ied Eco logy 40: 1060-1076. Lindroth, C . H . 1961. The Ground Beet les of C a n a d a and A l a s k a . Opuscu la Entomologica Supp lementum X X - X X I V . P p . 1-1192. Thie le, H .U. 1977. Carab id Beet les in their Envi ronments. Spr inger Ver lag , Ber l in. P p . xvi i-369. 114 6 4 Consumption of choice 17> Food choice Appendix Figure 1.1 Consumption of foods by Scaphinotus angusticollis. 115 

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}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            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.831.1-0075024/manifest

Comment

Related Items