8/15/2019 Anderson 1978

1/21

The Role of Aquatic Invertebrates in Processing of

Wood Debris in Coniferous Forest Streams

N. H. ANDERSON

J. R. SEDELL

L. M. ROBERTS

and

F. J. TRISKA

Reprinted from

THE AMERICAN MIDLAND NATURALIST

Vol. 100, No. 1, July, 1978, pp. 64-82University of Notre Dame Press

Notre Dame, Indiana

8/15/2019 Anderson 1978

2/21

8/15/2019 Anderson 1978

3/21

The Role of Aquatic Invertebrates in Processing ofWood Debris in Coniferous Forest Streams

N. H. ANDE RSONDepartment of Entomology, Oregon State University, Corvallis 97331

J. R. SED ELL, L.M. ROBER TS and F.J. T R I S K ADepartment of Fisheries and Wildlife, Oregon State University, Corvallis 97331

ABSTRACT: A study of the wood-associated invertebrates was undertaken in sevestreams of the Coast and Cascade Mountains of Oregon. The amount of wood debriswas determined in terms of both weight and surface area. Standing crop of wood perunit area d ecreases with increasing stream order.

Invertebrates associated with wood were functionally categorized and their bio-mass on wood determined. Major xylophagous species were the caddisflyHeteroplectrocalifornicum), the elmid beetle (Lora avara) and the snail (Oxytrema silicula). Staing crop of these species is greater on wood in the Coast Range than in the Cascades,which is attributed to species composition of available wood debris. The density ofavara was strongly correlated with the amount of wood available irrespective of streamsize within a drainage. The standing crop of invertebrates was about two orders of magnitude greater on leaf debris than on wood.

A potential strategy for wood consumption, based on microbial conditioning, is presented. The data are used to develop a general scheme of wood processing by inverte-brates in small stream ecosystems. Their impact is similar to that of invertebrates whichprocess leaf litter in terrestrial and aquatic environments when the full decompositioncycle of wood debris is considered.

I NTRODUCTIONT he allochthonous inputs to streams in western coniferous forest

erous need les, deciduous leaves an d woody ma terial, rangin g in sizan d bark to large logs. The amount of fallen wood in these streams large. Froehlich (1973) estimated that in one wa tershed of old -gr(Pseudotsuga menziesii) the stan din g crop of wood d ebris (pieces largediam) was more than 15 kg/m . The standing crop of small debris in the samestream was 1.08 kg/m (Sedell et al., 1974).

In view of the quantities of woody material in these streams, itwood has a significant role in en ergy flow, n utrient d yna mics, stran d in shaping the biotic community of these lotic ecosystems. A lthougists have emphasized the importance of allochthonous debris as thstream invertebrates, most previous studies are based on leaf inpuCumminset al., 1973; Bolinget al., 1974). Current literature on aqua tic communities inhabiting logs or inundated trees has emphasized ththese sites as habitats for attachmen t or surfaces for grazing of peri1968; N ilsen a nd L arimore, 1973; McLachlan , 1970) rather thanenergy an d nutrient inputs to the aqua tic system.

T he present study is a preliminary investigation of the wood comerous stream ecosystems of western O regon and of the role of invebiological processing or d egradation of wood. The objectives were toassociated with wood in streams an d to determine some of the intethe fauna and the wood substrate. In order to develop generalizations on theinvertebrate-wood in teractions, we chose to compare large an d smad ifferent areas ra ther than to investigate one site in d etail.

64

8/15/2019 Anderson 1978

4/21

Coast RangeSedimentaryTyee sandstoneUnconsolidatedSand and gravel360025 0

ca. 130

Cascade RangeVolcanicBreccias and uConsolidated-armoredB oulders, rubble and bo270022 0

ca. 450

1978 ANDERSON ET AL.: WOOD PROCESSING IN STREAMS 6SITE DESCRIPTIONSSeven streams, three in the Coa st Ra nge a nd four in the western Cascade R an ge, were sampled for wood an d associated invertebrates duPhysical and environmental data for the streams are provided in T ablof the streams stud ied, with the exception of B erry C reek, could bedraining dense coniferous forests, with a wet, mild climate in ruggedterrain.

Streams in both the Coast and Cascade ranges have fairly comparable watechemistry. ThepH ran ges from 6.8-7.4. T otal dissolved solids a re betw

TABLE 1.-Physical characteristics of the seven Oregon streams sampledfor wood and associated invertebrates

Stream

Strahlerstreamorder

Altitude(m)

Drainagearea

km2)Gradie

(%)Coast Ran ge

Berry Creek 1-2 75 3.22 1.7Flynn Creek 3 209 2.02 2.5Five Rivers 6 40 295.00 1.4

Cascade RangeDevil's Club Creek 1 835 0.05 35Mack C reek 3 830 5.35 20Lookout Creek 5 420 60.20 12McKenzie River 7 410 1642.00 9

TABLE I ( continued )

Discharge (ms)Summer Winter Annua lmean

StreawidtJuly

Coast RangeBerry Creek 0.005 0.03 0.02 2.0Flynn C reek 0.01 0.80 0.12 3.0Five R ivers 1.05 38.0 16.90 18.8

Cascade RangeD evil's Club Creek

8/15/2019 Anderson 1978

5/21

66 T H E A M E R IC A N M ID L A N D N AT U R A L IS mg/liter and total alkalinity is between 10-13 mg/liter for all theThe Cascade Range streams had higher phosphates (150-200 pg/14) thaof the Coast R an ge (40-60ttg/1 PO4), and lower nitrates ( < 10-502g/1 N Othe Coastal streams (170-1200pg11 NO3).C O A S T R A N G E S T R E A M S

F lynn Creek an d Five R ivers lie 16 km an d 35 km, respectiveO cean on the W slope of the Coast R an ge. In its na tural conddensely forested with douglas-fir and red alder(Alnus rubra). F lynn C reemain ed in this state an d is overgrown with salmonberry(Rubus spectabilis) amaple (Acer circinatum). La nd ad jacent to the sample site at F ive Rlogged and planted in pasture. A corridor of red alder and big-leaf maplemacrophyllum), approximately 25 m wide, borders the stream.

The area has a maritime climate with warm, dry summers and Prolonged periods below —6 C or above 38 C are uncommon . Msampling a reas is about 250 cm an nua lly, with 70% falling betwMarch (Hall and Lantz, 1969).

B erry Creek is on the eastern side of the Coast R an ge, approximthe Pacific O cean. A den se canopy of deciduous trees, primarilyleaf maple, covers the stream. T he area ha s topographically gentlfoothills form a bound ary with the Willamette Valley. T he hydrB erry Creek is similar to the Coa st Ra nge but with only a bout 1The study section of Berry Creek is a portion of the original streamcan be controlled by diverting excess water through a bypass chwinter flows are between .014m3/sec an d .028 m3/sec (Warrenet al., 1964).

C A S C A D E R A N G E S T R E A M ST he four streams studied in the Ca scade R an ge are located in

Andrews Experimental Forest about 65 km W of Eugene, Oregonrounding each of the streams are a dense mixture of douglas-fir, (Tsuga heterophylla) and western red cedar(Thuja plicata). A shrubby covine maple and red ald er borders the larger streams.

T he basic clima te of the area is a lso maritime, but with a grerange than in the Coast Mountain s. T emperature extremes ran ge during un usually cold winters to over 38 C for brief periods a lmMean annual temperatures are around 9.5 C, with a January mean of 2 C aJuly mean of nearly 20 C. Average annual precipitation for the per239 cm. Precipitation is strongly seasonal with 72 occurring from Novethrough March and only 7% from Jun e through September. In gewinter snowpacks can be expected above 1000-1200 m elevation; btions snow cover is erratic (Rothacheret al., 1967).

Streamflow of the three smaller streams is not well regulated.follows precipitation patterns very closely, with high flow during tan d an an nua l recession to low flow in late summer or early autuyears, peak flows are 1500-2000 times higher than summer low flow

The McKenzie River is well regulated because the porous lavas of theCascades store large quantities of snowmelt and release the water gfrom these lavas cause relatively high streamflows in late summestreams studied.

M E T H O D SAt each stream sampled, the study site consisted of a reach appro

the mean width. This reach was assumed to include all the major

8/15/2019 Anderson 1978

6/21

NATUR ALIST 100 1) 1978 ANDERSON ET AL.: WO O DPROCESSING IN STREAMS mg/ liter for all the streams sampled.ates (150-200J.A .g /1 PO4) than thoser nitrates ( < 10-50jAg/1 NO3) than

AMS35 km, respectively, from the Pacificits natural condition this area was

Alnus rubra). F lynn C reek has re-onberry(Rubus spectabilis) and vinesample site at F ive R ivers has been

d alder and big-leaf maple (Acerr s t h e s t r e a m dry summers an d cool, wet winters.

re uncommon. Mean rainfall at the70% falling between N ovember an d

st Range, approximately 85 km fromus trees, primarily red alder and big-ographically gentle terrain where thete Valley. The hydrologic regime ofwith only about 120 cm of ra infall.the original streambed in which flowrough a bypass channel. Controlled3/sec (Warrenet al., 1964).

REAMSan ge are located in or near the H. J.of Eugene, Oregon. The forests sur-ture of douglas-fir, western hemlockuja plicata). A shrubby corridor ofams.ime, but with a greater temperatureure extremes range from near –18 Cr brief periods almost every summer.with a January mean of 2 C and aipitation for the period 1952-1975 wasith 72% occurrin g from N ovembergh September. In general, permanent1200 m elevation; below these eleva-967).not well regulated. T he hydrograph high flow during the win ter monthsmmer or early autumn. D uring mostan summer low flows.cause the porous lavas of the H ighrelease the water gradually. D ischargelows in la te summer, unlike the other

sted of a reach approximately 20 timesnclude all the major habitats found in

the stream (riffles, pools, debris dams and alcoves). Sites were selecwithin each reach an d a ba nd transect was established to sample theBand transect lengths were related to stream size (1 m, Devil's CluBerry and Flynn creeks; and 5 m for Mack an d Lookout creeks andFive R ivers). Four to six samples were taken per stream.

WOOD DEBRIS ESTIMATESWood was either wet-weighed or measured for conversion to d ry we

face area. La rge bran ch an d bole wood ( > 10 cm diam) were measurd iameter of the small an d large end , an d length. Measurements wervolume per piece and volume per area of stream accord ing to the technby Lammel (1972) from Brown (1971) and Van Wagner (1968). Volume wasfurther converted to weight and surface area. In each band transecdebris ( < 10 cm diam) was collected by hand and placed in buckets orcollected the wood debris from deep areas. The surface area of the smderived from individual measurements of several hundred sticks an( < 10 cm d iam) from L ookout Creek. This derived constan t of 0.130 m /kg of wowas applied to wood from the other streams. Average diameter of thisappea red compa rable from stream to stream . T he surface area estimconservative since each piece of wood debris is treated as a cylinder or t

A djustments were made on F ive Rivers and the McKenzie River tcentral channel which was practically devoid of wood and could not befaunal area. On Five Rivers, the faunal area comprised 40% of the totThe faunal area comprised 15% of the total stream area of the McKenwas restricted to a 2 -3 m ban d at the river's edge. T he entire streambestreams was considered to be a faun al area.

INVERTEBRATE COLLECTIONSA pproximately 200 kg (dry weight) of sma ll wood d ebris was ex

field. The larger insects and snails were removed with forceps. After stick in a tub of water, the resulting debris was retained by a 0.5-mm mein alcohol and sorted at the laboratory. The beetle,Lara avara (Elmidae), andcaddisfly, Heteroplectron californicum (Calamoceratidae), and other dominantgroups were counted an d weighed ind ividually.

With the exception of theLara avara larvae, the invertebrates were driedfor 24 hr and weighed. The L. avara were live-weighed (the larvae wereother experiments) an d dry weights were obtained from a regression eq— 1.18 -I- 1.04(X) ; X = wet weight; R = .87; n = 40). Tissue weights of snailswere obtained empirically, resulting in a value of 20% of dry weight i(D . McC ullough, pers. comm.). In sect weights, exceptL. avara, were increase25% to ad just for weight loss in alcohol (Ma ckay a nd Ka lf, 1969).

SOURCES OF SAMPLING ERRORThe estimates of invertebrate numbers and biomass are approxima

allow comparison between streams. T hese are und erestimates because sresulted in losing or missing part of the population.

Some invertebrates were lost or escaped when the wood was picked uwas greatest in deep areas where diving wa s required a nd in fast watebe greatest for vagile species that use the substrate primarily as a rest(e.g.,Baetis spp.) and forHeteroplectron californicum, which is easily dislodged .

Sma ll twigs and other woody d ebris (less than 5 mm diam) were undin collections. This size class has a large surface-to-volume ratio andwhere it is impossible to distinguish the wood habitat from a core sample

8/15/2019 Anderson 1978

7/21

68 T H E A M E R IC A N M ID L A N D N A T U R A L IS T 100When picking out the wood an d associated faun a, the ind ividuals ban d dislodged to such an extent that there was n o way of d eterminilocation. S omeHeteroplectron californicum, with their stick cases, were missed in this way.N o satisfactory technique was d eveloped for sampling the very s

at the other end of the scale, logs larger than 40-cm diam. Attemptexamine the logs and to pick invertebrates under the water, but these were o

qualitative collections. C on siderable effort was required to removresultan t disturbance caused losses of an unkn own portion of the faunis lodged in debris jams, removal of key sticks caused a flushing acaway small wood a nd part of the fauna .

Visual examination of woody material was time-consuming, tedious andaccurate. Loca ting ind ividuals on this substrate was a lso greatly influcond itions. T hus, coun ts on rain y da ys or those in dimly lit forestedFinding the insects on the wood substrate is difficult because species ahabitats (e.g., some stoneflies, mayflies and chironomids) were ableor beneath bark, while others, such asLara avara, are cryptically colored, seasily overlooked.

R E S U LT S A N D D I S C U S S IO N

Q U A N T I TIE S O F W O O D D E B R I S IN T H E S T R E

The quantities of wood found in the streams (Table 3) reflect bobetween the Coast Range and Cascade Range, as well as a small-to-large strpattern of decreasing amounts of wood.

Coa st Ra nge streams contain little large bran ch or bole wood cCa scad e streams studied. T his is probably due to two factors: (1)trees of the former are about 130 years old and quite healthy, and wood coming into the stream is alder which is comparatively smalldecomposes quickly. T he forests in the Cascad es are much older (ca.ma ny trees bad ly affected with heartrot and more susceptible to wconifer trees falling into the streams in the Cascades are typically ovewhile most of the alders an d other trees in the Coa stal streams a re lesdiam. T he smaller the wood diameter is, the easier itis to be moved by the hyproperties of the stream.

TABLE 3.-Estimated quantities and surface areas of wood debris in seven Oregon streams

Large branch and bole wood (>10 cm)

Small branch wood' (

8/15/2019 Anderson 1978

8/21

N AT U R A L I ST 100 1) 9 8 A NDERSON ET AL.: W OO P ROCESSING IN STREAMS,a , the ind ividuals became d ispersedo way of determining their originalth their stick cases, were probablyampling the very small material or,-cm diam. A ttempts were made toer the water, but these were onlyrequired to remove logs, and then portion of the fauna . Where woodaused a flushing action that washed

s time-consuming, tedious an d in-as also greatly influenced by lightingn d imly lit forested a reas were low.ult because species adapted for thesenomids) were able to hide in cracks, are cryptically colored, sessile and

USSION

THE STREAMSms (Table 3) reflect both differences

, as well as a small-to-large streamnch or bole wood compared to the

o two factors: (1) the coniferousd quite healthy, and (2) most of thecomparatively small in diameter andare much older (ca. 450 years) with

more susceptible to windthrow. Thead es are typically over 50 cm in diam,oa stal streams are less than 25 cm inier it is to be moved by the hydraulic

wood debris in sevenOregon streamscm)

Sma ll bran ch woods (

8/15/2019 Anderson 1978

9/21

70 T H E A M E R IC A N M I D L A N D N AT U R A L IS T 100(1)when p laced on unsuitable wood or if they are stressed by high tempedissolved oxygen.Larvae have been held at 15 C and long day-length (16 1: 8 d) in shallowdishes of water for several months. Under these conditions the larvaenormal and chew grooves in stream-collected sticks to establish a typicaThe larvae produced considerable feces (up to 10-30% of their body weiHowever, on the basis of wet weight comparisons over 5 months, they did

in weight, so aspects of the rearing procedure need to be modified to procond itions for growth.H eteroplectron californicum.—T he larvae of this caddisfly are dist

they hollow out twigs or pieces of wood for their portable cases. A fewalso been found in wood, 2 cm diam and 15 cm long, which obviouslytransported. A lthough the larvae have been shown to be effective shredin laboratory studies, their characteristic attachmen t to wood suggests may be placed in the xylophagous category.

Win terbourn (1971) suggested that the life cycle ofHeteroplectron californicuminvolved a rapid growth period during the summer because small larlected in June or July an d fina l instars were most abunda nt in A ugust H owever, our da ta ind icate a 2-year life cycle without any periods o(Anderson, 1976) . Midwinter collections were composed of two mainthird in star larvae (mean dry weight, 0.94 mg) an d fifth instar larvae5-40 mg. A s the emergence period of adults is relatively short, ran ging

to mid-July, it seems likely that the overwintering final instars represenalread y in its 2nd year that would pupate in the sprin g, whereas thewould require the next season to complete development. A lso, collectiMay to early July (during the adult emergence period) contained aboinstars and these would require another year to reach maturity.

Oxytrema silicula.—This snail is the most obvious invertebrate in westreams, being particularly abundant in the Coast Range and Willamette ValleyBenthos densities can exceed 3000/m , resulting in a tissue biomass of 3-5 g/m . Thesna ils occur on various substrates in both riffles and pools.

The life cycle ofOxytrema silicula extends over several years, so a rangclasses occurs in all seasons. Eggs are laid in the spring and early sumemerged snails are abundant by July. There is a tendency for larger occur in larger streams. W here populations a re extremely high, the meto be low.

Oxytrema silicula is an extreme trophic generalist. It is an important speriphyton where primary prod uction is high as well as a d omina n t fallen leaves. The rasping method of feeding results in obvious skeletonizing odeciduous leaves. Whether the leaves were eroded away at the surface or eatencompletely through in isolated locations is not known. The characterilocomotion and feeding by snails explains their occurrence on flat or smsuch as rocks and leaves, and may also account for their presence assume that they are primarily grazing periphyton on this substrate, buaction of their radula a lso removes the superficial layers of wood.

FA U N A L S U RV E YField observations and literature data were used to categorize the

debris by larvae of various insect groups (Table 4) . The general nature of thisclassification reflects the lack of d etailed kn owledge of the natural hiaqua tic insects. The surface area a nd the large number of protective nicafford considerable living space an d con cealment. Wood is used for ov

8/15/2019 Anderson 1978

10/21k ft 4

Vs 11

1978 ANDERSON ET AL.: WOOD PROCESSING IN STREAMSTABLE 4.—Insects associated with wood in Oregon streamsColeopteraEhnidae

Lara avara: feeding (gouging), habitatPsephenidae

Acneus: scrapingOedemeridae

Uncommon but

Ditylus: tunnelling

restricted habitat

AmphizoidaeAmphiroa: habitat

TipulidaeLipsothrix: tunnelling

RhagionidaeAtherix: habitat, feeding

SimulidaeSimulium: attachment

ChironomidaeMany genera (diverse associations ranging from microhabitat specificity to chance settling)

EphemeropteraLeptophlebiidae

Paraleptophlebia: habitat, scraping (?)Heptageniidae

Cinygma integrum: habitat, scraping (?)Ephemerellidae

Ephemerella flavilinea (and other spp.): habitat, scraping (?)

Plecoptera

TrichopteraCalamoceratidae

Heteroplectron californicum: boring, gouging, pupa tingLepidostomatidaeLepidostoma unicolor (log cabin case-maker) : case construction, scraping and chewing,pupatingL. quercina (and other chimney case-makers) : scraping, attachment, penetrating forpupation

BrachycentridaeBrachycentrus: attachment, surface pupa tionMicrasema: scraping, penetration for pupation

LimnephilidaeHydatophylax Psychoglypha Halesochila Onocosmoecus, Cryptochia (wood case-makers):case-construction, shredding( ?), surface and some penetration for pupationNeophylax, Dicosmoecus, Allocosmoecus (stone case-ma kers) : surface pupationClossosomatidaeGlossosomaand Anagapetus: surface pupationHydropsychidaeArctopsyche, Parapsyche, Hydropsyche: net and retreat attachments, some pupationPolycentrododidaePolycentrapus: retreat site

PhilopotamidaeWormaldia, Dolophilodes: net attachments441tracr.philidaeR hyacophila (esp. acrapedes—group) ; habitat for predation (and detrital feeding?)

44, IN „It

m • N • I. II NN • • • •

•

N ♦ • • • •

lk l

e uses of woodature of this

history of theseniches on wviposition, as a

7 1

Diptera

PeltoperlidaePeltoperla: habitat, shredding ( ?)

NemouridaeNemoura s.1.) : habitat, shredding ( ?)

PteronarcidaePteronarcys, Pteronarcella: habitat, shredding (?)

ChloroperlidaeAlloperla: habitat, feeding

100 1

ure or low

shallowppear quitefeeding site.ht per clay).not increase

vide suitablective in thatlarvae haveould n ot beers of leavesat they also

alifarnicuma e were col-

d September.apid growthsize classes:anging fromom late Ma y

sent a cohortthird in starsons from latet 75% fourth

stern O regonmette Valley.

5 g/m . These

range of sizemer and newlyind ividuals toean size tend s

ant scraper ofconsumer ofeletonizing offace or eatenristic mode of

mooth surfacese on wood. Weut the rasping

8/15/2019 Anderson 1978

11/21

72 THE AMERICAN MIDLAND NATURALIST 10nursery area for early in stars, for resting, molting, pupation an d emeof its unique capillary properties, it affords a n ideal a ir-water interfaof temperature an d moisture.

Trichoptera.—The most conspicuous and diverse insects on wood wlarvae. They are relatively large and easily seen because of their cAll major types of caddisflies, free-living, net spinners and case-makedebris.

F ree-living larvae ofRhyacophila (especially theacropedes group) are coon wood surfaces.Rhyacophila spp. are generally considered to be predMecom(1972) listsR. acropedes as a d etritivore. It is probable that larvaof the rare species in this very diverse genus will be found that are sdebris microhabitats.

Several genera of net-spinn ing larvae are associated with debrimerged wood in lotic habitats. E xploitation of the habitat for net anments is largely opportunistic, but there may also be a degree of specificity. example,Polycentropus halidus is uncommon in these O regon streams, buthe larvae an d their nets were found within pieces of decaying wood.

Among the case-making larvae, the uses of the wood substrate may be nospecific, such as larval or pupal case-attachment, or they may be speciin selection of case material or in food intake.Cryptochia andPedomoecus, obgenera of western limnephilids, were collected d uring this survey; tha reflection of microhabitat specificity. In the McKenzie River, thespecies collected on wood wa sBrachycentrus americanus. H owever, removal odebris would n ot greatly affectB. americanus as the larvae can attach theirstones just as well.

In add ition toHeteroplectron californicum, other caddisflies intimately awith wood d ebris include severalLepidostoma spp. and a number of genera cased limnephilids. P resumably because of their close association wthey also have the greatest impact in degradation. These species are pin the shredder functional feeding group. We suggest that, in additioleaves, all of these also feed to some degree on wood an d the associ

Limnephilid gen era that build wood or bark cases, such asHydatophylax, Psycho-glypha, Onocosmoecus andHalesochila, are amon g the largest insects foundin O regon streams. In ad dition to their feeding activities, their cahavior has an impact on the particle size of woody debris.Lepidostoma spp. use twigs and pieces of bark in case construction. Field observations indspend considerable time attached to wood , either feedin g or restinLepidostomunicolor larvae were observed to remove the bark of small twigs by chespecies ofLepidostoma, including both wood-cased a nd san d-cased spwithin cracks in wood. In soft wood they may chew a hole for a pup

C'oleoptera.—Although several families of beetle larvae were collectenon e was common exceptLara avara. T he psephenid , oedemerid a nd abeetles were each represen ted by less than five ind ividua ls but theyTable 4 because they are not commonly collected and records of thmay lead to further investigation.

Diptera.—On a n umerical basis, this order was the most abund an t in the wood samples. Chironomid larvae were ubiquitous, but it is diffize about the role of this family because of the great va riety of specT he tubes ofRheatanytarsus were frequently abunda nt on wood a lthougso than on adjacent rocks, so we do not ascribe any preference for thas a surface for attachment. An un described species ofBrillia (det. W. P. Cof

was collected un der the bark of recently fallen douglas-fir branches.

8/15/2019 Anderson 1978

12/21

ets

lys.d

nutmeod

b-h-

orof

n-le,urebeant

dyto

tedod-tat,dedingora.ho-

oodbe-

mallvaema

eralate

ood,zoidd initat

ectseral-bits.

morebitat

man)were

1978 ANDERSON ET AL.: WO O DPR O C E S S I N G I NST R E A M S 73tunnelling in the live phloem and outer cortex of the wood. This species is a colonizerin the first stage of degradation, in contrast to other species of this genus that areknown to bore into soft, waterlogged wood (Coffman,1978) .Larvae of the tipulid,Lipsothrix spp. were common in Berry Creek. These slender,

wormlike larvae occurred in tunnels in decayed alder wood that was so soft it couldbe broken apart by hand. Rogers and Byers(1956) provide detailed observations ofL. sylvia in small A ppalachian streams that demon strate the restriction of this speciesto soft, waterlogged wood.

Ephemeroptera.—Although a n umber of mayfly species were collected on wood,most of these are considered chan ce associates(e.g., Baetis spp.) that are basicallybenthic species. Some taxa occurred with enough consistency to suggest a real a ssocia-tion. I n D evil's Club Creek,Paraleptophlebia pallipes is a characteristic species onwood, while in other sitesP. debilis andP. temporalis occurred on wood a s well as inleaf detritus of backwaters and marginal areas. Among the heptageneid mayflies,Cinygma integrum was largely restricted to wood substrates, where it apparentlyingests periphyton a nd woody d etritus by scraping.Epeorus nitidus is potentiallyanother wood-associated species as it is found in the wood-choked first order streams.Severa l species ofEphemerella were common in the samples—especially from thelarger streams—E.flavilinea appea ring to be the most chara cteristic species of thewood habitat. Further study may well demonstrate some feeding specializations withinthis genus for exploiting the wood debris habitat.Ameletus spp. appeared to use woodpreferentially as a site for emergence. T his genus is somewhat un usual for mayfliesin that larvae leave the water prior to moulting to the subimago.

Plecoptera.—The most common stoneflies in the collections were the detritivores,Nemoura and Peltoperla. Pteronarcids were less common but their large size madesingle individuals conspicuous.Leuctra spp. occurred on wood in sma ll streams; theirslend er bodies are well-ad apted for penetrating into the cracks an d crevices.Alloperlaspp. were also common on wood. T hese preda tors were presumably using this sub-strate as a source for prey.

U A N T I TAT I VE E S T I M AT E S O F W O O D - A S SO C I AT E D I N V E RT E B R AT E SThe results of the extensive sampling program for the seven streams are given in

Table5. The biomass of the dominant xylophagous insects(Lara avara andHetero-plectron californicum) is compared with that of the other wood-a ssociated insectsand snails. Invertebrate biomass is expressed both on a wood-weight and on a surface-area basis. The ma jority of organ isms collected were from wood of intermediate size(ca. 1-10 cm diam) .

The data indicate a greater standing crop of invertebrates on wood in streams ofthe Coast Range than in the Cascades. Although this could be attributed to differencesTABLE 5.—Biomass of invertebrates on small wood debris in seven Oregon streams, July 1976

Invertebrate biomass/weightof wood (mg/kg)

LaraHetero-plectron

Otherinsects Gastropods

Coast RangeBerry Creek 69.8 6.7 110.0 710.9Flynn Creek 24.3 19.8 45.6 91.6Five Rivers 19.8 0 23.2 72.9

Cascade RangeDevil's Club Creek 12.9 0 14.8 0Mack Creek 13.5 3.6 46.2 0.1Lookout Creek 15.2 4.4 109.8 24.3McKenzie River 19.3 0.1 62.0 36.4

8/15/2019 Anderson 1978

13/21

7 4 THE AMERICAN MIDLAND NATURALIST 100 1TA BLE 5.—( continued)Invertebrate biomass/surface

area of wood (mg/m2)

LaraHetero-plectron

Otherinsects Gastropo

Coast RangeBerry Creek 538.7 51.9 848.9 5486.6Flynn Creek 187.6 152.8 351.7 706.9Five R ivers 152.7 0 179.4 562.9

Cascade R angeDevil's Club Creek 100.0 0 113.9 0Mack Creek 104.5 27.9 356.4 0.7Lookout Creek 117.1 34.2 845.9 186.8McKenzie River 148.7 0.9 478.3 280.6

in environmen tal factors (Ta ble 2), we a ttribute the higher d ensity of invthe Coast Range streams to the presence of more alder as a substrate and The data for colonization rates on wood in Berry Creek support this(Table 9) . The density ofLara avara is strongly correlated with the amouspecies of wood av ailable, irrespective of stream size within a d rainadensity ofHeteroplectron californicum, however, is more restricted by streamThis species was essentially absent from large streams in both areas athe first-order D evil's Club Creek. A bsence from the latter 'was unexpl

habitat of this species is known to include some very small streams (And1976 ) .

The data for "other insects" is more variable than for the xylophagouswas expected because many of the individuals may only be chance assoand Larimore, 1 9 7 3 ) . In addition, limitations of the collecting techniques alsoincrease the variability of these da ta. T he biomass of the "other insect" gfrom equal to that of the xylophagous group(Lara plusHeteroplectron) in D eviClub Creek and the Coast Range streams, about 2.5-5.5 times greater inLookout Creek and the McKenzie R iver.

The snail, Oxytrema silicula, was the domin an t invertebrate by biomaswood samples from Coast Ra nge streams. In the Cascades, snails werMcKenzie R iver and L ookout Creek but were largely absent a t the altiCreek and Devil's Club Creek. Since snails consume a wide variety of fhigh biomass, their impact on the wood substrates may be greater thanwood-feeding in sects.

The fauna of the experimental section of Berry Creek is atypical Coast Range stream because freshets are diverted through a bypass prevents flushing by winter storms. A s this site has not been flushed for oa considerable amount of organic matter has accumulated. This lack ofaccount for the high density of insects an d the extraordin ary biomass of of the organic accumulation in Berry Creek is small twigs and fine particles notsampled as wood.

In addition to the xylophagous feeding group, some invertebrates obment from the wood substrates by grazing or scraping the autotrophic McLachlan( 1 9 7 0 ) and N ilsen an d L arimore( 1 9 7 3 ) suggest that the large numof chironomids responded to the development of algal films or mats, but measure the quantity of periphyton. Table6 compares stand ing crop of periphmeasured by chlorophyll extraction, of wood with that on stream rocksCa scade R an ge streams. In the densely shad ed first-order stream, periphis low on both substrates. In the other streams, the biomass on wood is clower than on stream rocks. H owever, the amount of periphyton on wo

8/15/2019 Anderson 1978

14/21

ANDERSON ET AL.: WO O DPROCESSING IN STREAMS 75ad equate to support a low population of grazers. Periphyton ingested by sn ails, whenthey rasp the wood surface, probably contributes significantly to their nutrient intakebecause of the higher food qua lity. T he feeding activity also exposes additiona l sitesfor microbes. T his ind irect role in the processing of wood in streams is little und er-stood an d is probably greatly und erestimated.

COMPARISON OF WOODY DEBRIS FAUNA WITH LEAF-PACK FAUNAA lthough there is a d iverse fauna a ssociated with wood, the biomass is low. T otal

invertebrate biomass (mg/kg of wood) ranged from about28 in Devil's Club Creekto 897 in Berry Creek (Table5 ) . If the latter value is excluded as being exceptionaldue to lack of winter flushing, then the maximum biomass wa s181 mg/kg for FlynnCreek.

T he relative paucity of the wood fauna is strikingly illustrated by comparin g thestanding crop with that obtained in leaf detritus in Flynn and Mack creeks(Table 7 ) . T he leaf-pack method reported by Petersen a nd Cum mins(1974) wasused. T he da ta for leaves are a composite mean of invertebrate biomass per gramleaf pack of a lder, vine ma ple, big-leaf ma ple and douglas-fir leaves attached to bricksto simulate na tural leaf accumulations. T he packs were placed in the streams d uringN ovember an d sampled at a bout monthly intervals until May (S edellet al., 1975).

T he stan din g crop of each group (shredders on leaves, gougers on wood, totalinsects an d sn ails) is about two orders of magnitude greater on leaves than on wood.Underestimates of the wood fauna could account for some of the difference but evenif the biomass on w ood w ere doubled, the comparison would rema in essentiallyunchanged. The seasonal difference in the comparison (leaf packs from November-May, wood sampling in July) undoubtedly contributes to some of the wide divergencein va lues. H owever, amon g the dominan t componen ts on both substrates, the snailbiomass does not change greatly on a seasonal basis, and shredders such asLepi-dostoma unicolor, Hydatophylax hesperus and Onacosmoecus sp. are approaching

TABLE 6.—Comparison of periphyton standing crop on wood' and mineral substrates2in four Cascade Range streams in the same drainage system

Strahlerstream

Woodsurface

Mineralsubstrate

Stream order g/m2 g/m2

Devil's Club Creek 0.2 0.2Mack Creek 3 0.7 1.5Lookout Creek 5 0.4 5.5McKenzie River 0.6 5.0

Analysis of periphyton was from small sticks ca. 0.5-2.0 cm diameter and 10-25 cm long(Naiman and Gregory, pers. comm.)

2 Lyford and Gregory (1975)

TABLE 7.—Comparison of standing crop of invertebrates in leaf packs and wood from aCoast Range and Cascade Range stream of similar size. The dominant processors on leaf packsare shredders, on wood are gougers

Total Totalinsects Gastropods invertebratesmg/kg mg/kg mg/kgFlynn Creek

Leaf packsWood

Mack CreekLeaf packsWood

1978

16 960

63.4

9 01089.7

8 80263.3

7,95091.6

00.1

Dominantprocessors

Mg/kg

6 02044.1

4 80317.2

181.3

8 802

8/15/2019 Anderson 1978

15/21

76 T H E A M E R IC A N M ID L A N D N AT U R A L IS T 10maximum weight during the summer (An derson, 1976) .The difference in invertebrate biomass on leaves an d wood is a ttrito differences in food quality. Although both are low in n itrogen cperiphyton, seed s or fresh macrophytes (Ta ble 8), the wood is so higtory components (lignin and cellulose) compared with leaves, that it becomesavailable at a very slow rate. The greater surface area and penetrability of learesults in microbial conditioning occurring within months, compared

wood. C ond itioning is a key factor in the d ebris becoming av ailableinvertebrates.

I M PA C T O F IN V E RT E B R AT E S O N W O O D D E G R A DLaboratory studies were und ertaken to obtain prelimina ry estimate

of invertebrates on degradation or reduction in particle size of woodwere held at 15.6 C in dechlorinated tap water in shallow dishes or pans. Thequan tity of feces produced is useful as an ind ex of food con sumptionLara avara larvae (N = 20) produced 0.15 ± 0.05 mg feces/mg body wtstream-conditioned wood compared with 0.03 ± 0.01 mg/mg/d ay wiilized with ethylene oxide, ind icating a feed ing response to the presemicrobial flora. However, even with high fecal production rates byL. avara, we hnot demonstrated growth under laboratory conditions, possibly becauseflora changes after a period of time in the laboratory.

Fecal collections were also used as an index of food consumptHeterplectron cakornicum larva e supplied with va rious foods (T able 9). Wittion of the "twigs +Quercus leaves" series, the larvae were fourth or newfifth instars. Fecal production was about twice as high for larvae fedtioned wood (big-leaf ma ple) compared with those fed a lder leavewere readily consumed; the larvae chewed a hole in the husk and fewithin a few hours after the wheat was placed in water. The amounduced on the wheat diet was the lowest for any of the diets. This malowered con sumption an d higher assimilation efficiency of wheat graidiets the larvae were observed attached to and chewing on cases of

TABLE 8.-Nitrogen concentrations and carbon quality of selected particulateinputs to coniferous forest streams

N (%)C/NRatio

Cellulose(%)

Lignin(%)

Total fib(celluloselignin)(%)

Wood (douglas-fir)Twigs (with bark) .20 235 40.2 34.4 70.6Bark .15 324 48.0 9.9 56.9Wood .04 1343 31.7 47.6 79.3

LeavesRed alder 2.03 23 9.0 9.5 18.5D ouglas-fir .51 97 14.5 24.2 38.7Big-leaf maple .74 62 16.3 17.3 33.6Vine maple .56 77 14.7 8.5 23.2

OtherMosses' 0.8-1.2Algae 6-10Flowers and fruit parts 3 1-2

1 J. Lyford (pen. comm.)2 S. Gregory (pen. comm.)3 Cromack and Monk (1975)

8/15/2019 Anderson 1978

16/21

1 0 7 8 ANDERSON ET AL.: WOOD PROCESSING IN STREAMS

the dish. Thus, wood was an ingested component in all trials. The larvae egested1-2 times their body weight per day of fine particle feces with wood as food.

Field studies of invertebrate colonization were conducted by placing 46 woodsticks alder (14), douglas-fir (14) and hemlock (18)-in Berry Creek duringApril 1975. These were kiln-dried commercial grade 1 X 1 (2.5 cm X 2.5 cm)lumber, 0.92 m long, that were grooved with a saw blade 2 mm wide and 2.5 mmdeep. The sticks were examined every 2-3 weeks for 56 weeks. The number of Laraavara larvae was recorded without removing them from the sticks.

Of the 232 Lara avara recorded between weeks 2 and 56, 44% were on alder,38% on hemlock and only 18% on douglas-fir (Table 10) . Half of the records areconsidered to be repeat counts of the same individuals at successive sampling intervals.For example, one larva remained in the same position for over 4 months. Larvaewere recorded as arrivals when first counted and as residents when there was alarva on the same stick at two successive sampling dates. The number recorded as arrivals (Table 10) indicates that similar proportions of larvae initially landed oneach type of wood. However, the residency index shows that the percent remainingon douglas-fir for 2 or more samples was less than half of that on hemlock or alder.That the grooved wood is attractive to drifting L. avara is apparent in that a similaramount of alder and douglas-fir branch material, placed in the stream at the sametime as the sticks, only attracted six larvae compared with the 232 on the sticks.McLachlan (1970) demonstrated that submerged wood previously damaged by barkbeetles was more readily colonized by mayflies than trees with solid bark.

After 15 months in the stream, alder showed a dramatic deterioration, whereas

douglas-fir appeared unchanged and hemlock was intermediate (Fig. 1). The surfaceof the alder was abraded to the extent that the saw grooves were no longer apparent.The numbers of Lara avara alone were too low to have produced this change, basedon laboratory studies of feeding rates. Field observation and collection indicated thatthe feeding activity of the entire invertebrate complex was involved, and that snailswere the major component. The mean numbers of biomass of Oxytrema silicula perm 2 of wood in August 1976 were: alder, 91, 890 mg; hemlock, 121, 770 mg; anddouglas-fir, 43, 279 mg. In contrast, the numbers and biomass of L. avara were:alder, 7, 45 mg; hemlock, 5, 35 mg; and douglas-fir, 3, 20 mg.

TABLE 9.-Fecal production by Heteroplectron californicum larvae ondifferent food substrates at 15.6 C in the laboratory

FoodN o. oflarvae Days

Totalfeces(mg)

Terminal

Fecal

wt. of larvae production(x)

(mg/mg/day)

Stream wood Acer) 9 13 354.15 1.49 2.15Stream wood(Acer) 59 9 1323.15 2.27 1.10Stream wood (Acer) 10 21 826.85 4.09 1.01Alnus leaves 9 13 224.95 3.90 0.52Alnus leaves 10 21 404.83 4.08 0.47Wheat grains 10 20 363.50 4.92 0.37Twigs + Quercus leaves 26 6 219.02 9.58 0.15

TABLE 10.-Colonization of three species of wood by Lara avara larvae inBerry Creek, B enton C o., Oregon, 1975-76

Alnus

Tsuga

Pseudotsugarubrci h eterophylla menziesii Total

No. of larvae recorded (weeks 2-56) 101 89 42 232N o. counted only once (Arrivals) 49 36 31 116No. recounted at next sampling interval (Residents) 52 53 1 1 116"R esiden cy" Index (Percent on each wood) 51% 60% 26%

.* k

t•

N \ %■ * N,

c 1,1k

^‘

111/4,

N C 0 1 ,

.4k

* N

N

r

‘0, • 1 lk

11

14

Is 14.

N,

• • • • .•It v. k qk

*19*1/4 1k.

77

44 X.‘ 3 r o

41

8/15/2019 Anderson 1978

17/21

78 THE AMERICAN MIDLAND NATURALIST 10

Mil lllllllllI U M W 1111111111111111111111111111111111611111111111,11111111

SYNTHESIS

ECOSYSTEM ROLE OF WOOD AND LEAF LITTERBranch and bole wood in streams influences channel morphology

a trap for organ ic material and a break in the chann el grad ient. Dto retain materials in the stream lon g enough to be processed by botinvertebrates. T he wood d ams a lso produce habitats such as pools, madicolous areas (Hynes, 1970, p. 409), and by acting to dissipate the energrunning water, they decrease the amount of energy available to erode1976).

Woody debris and leaves, the two major allochthonous componstream, operate in d ifferent wa ys in relation to qua ntity, quality an(Table 11). The leaves form a small pool of readily available organthe wood forms a large pool of less available (slow turnover) orgapared to the leaves. The slowly processed wood also constitutes a loof essential nutrients. The composition, metabolic structure and nutime of the particulate organic pool effectively provides both a flexibiwithin the system.

STRATEGIES OF WOOD-ASSOCIATED RESOURCE UTILIZATIONInvertebrate growth and survival require a source of fixed energ

particularly n itrogen. T hese requiremen ts presen t special problemsuming invertebrates since most of the carbon is tied up in wood f

METRIC 1

Pseudotsuga

Tsu ga Ainos

Fig. 1.—Appearance of grooved wood of douglas-fir (Pseudotsuga), hemlock (Tsuga)alder (Alnus) after 15 months in Berry Creek. Note that the grooves in alder have almost dis-appeared

8/15/2019 Anderson 1978

18/21

1978 AND ERSON E T AL.: WOOD PRO CESSING IN STREAMS 79cellulose), and nitrogen concentrations are extremely low (Table 8) . Two basicstrategies may be involved in wood exploitation. One is the consumption of volumi-nous amounts of wood tissue to obtain enough digestible carbon and nitrogen fromboth the wood substrate and associated microbial flora to fill the energy and nutrientrequirements of the organisms. This strategy is common to leaf-shredding inverte-brates and appears to be employed by Heteroplectron californicum larvae which mayconsume up to 200% of their body weight per day of wood tissue in laboratoryculture. A second strategy is cultivation and retention of a gut flora to furnish

essential vitamins and amino acids and aid in the internal digestion of wood fiber(e.g., termites). A final strategy would be some combination of the abovemechanisms.

Despite the mechanism employed, strategies for wood exploitation are compli-cated in aquatic habitats by the waterlogged condition of the wood substrate.Microbial colonizaion of submerged sticks and twigs is primarily a surface phe-nomena (Savory, 1954), although some ascomycete and basidiomycete fungi whichcolonize wood prior to entry into the water survive and have been observed to producefruiting bodies on submerged and waterlogged wood. In the terrestrial environment,fungal colonization occurs throughout the wood matrix. Thus, microbial-invertebrateinteractions are possible for carpenter ants, beetles and termites which bore throughthe wood. In aquatic habitats, Lara avara produces only shallow gouges along thesurface not much deeper than the larva itself. It is only in the later stages of decom-position, when the wood is soft and punky, that species such as Lipsothrix occurdeep in the wood matrix.

In terrestrial habitats, wood consumption by invertebrates has been related tofixation of nitrogen in the gut and to other special microbial-invertebrate interactions.Nitrogenase activity by bacteria in the gut of termites (Breznak et al., 1973; Bene-mann, 1973) and cellulase activity by protozoa (Mannesmann, 1972) and bacteria(French, 1975) have been demonstrated as strategies for fiber utilization.

In aquatic systems, Carpenter and Culliney (1975) demonstrated both nitrogenfixation and cellulase digestion by microflora of shipworms in the marine environ-ment. Similar experiments have thus far yielded negative results with Lara avaralarvae maintained in laboratory cultures for several weeks. Gut flora of laboratory-maintained specimens have also been placed in the selective media of Hino andWilson (1958) for isolation of enteric nitrogen-fixing bacteria, and on cellulose media(Sigma MN-300) . The results of these experiments have also been negative. M. J.Klug (pers. comm.) has found little evidence for a resident gut flora by microscopicexamination. To what extent the result is a function of laboratory-rearing practicescannot be determined at this time.

TABLE 11.—Comparison of leaves and wood debris as components ofsmall coniferous forest stream ecosystems

Leaves Wood debrisStanding crop (% of particulateorganic matter)Seasonality of inputsImpact on stream morphologyCarbon:nitrogen ratioTotal fiber (%)Degradation timeMicrobial colonization patternInvertebrate componentInvertebrate utilization

60 (excluding boles)Erratic

Major (stabilize and destabilize)300-1000:1

70-805-200 years

Primarily surfaceLow (mg/kg)

Shelter, substrate, oviposition,pupation, emergence, food(direct and indirect)

8/15/2019 Anderson 1978

19/21

80 T H E A M E R IC A N M ID L A N D N A T U R A L IS T 10O ne alterna tive hypothesis to explain the feed ing strategy ofLara avara iconsumption of a thin surface of partially degraded wood and microal source of carbon a nd nitrogen, augmen ted by n onsymbiotic miIn itially this strategy did not seem feasible, sinceL. avara spends over 2 yelarva and consumes only 10-30 of its body weight/day on wood; only a smfraction of this diet is actual microbial biomass. Our preliminary sdecomposition in water, however, lend some support for this hypothefine wood debris in streams are colonized by both cellulose- an d lignwithin 6 months. Within 1 year,Alnus wood was softened to a depth of 0ad dition, n itrogen fixation was also observed within 3 mon ths by fron various wood substrates. Although the eviden ce is scan t,L. avara with a sttube gut andno current eviden ce of symbiotic gut flora (M. J. K lug, ma y be able to grow with the activity of cellulase, ligna se an d n isurfaces of wood providing the digestible material. Use of wood as would decrease competition with most leaf-litter shredd ers. H owevimposes a growth rate regulated by the rate of conversion of refractn itrogen to a digestible resource by the en zymatic activity of free-livfungi. T he energetic cost of wood exploitation is a long life cycle anmetabolic rate.

C O N T R IB U T IO N O F I N V E RT E B RAT E S TO W O O D D EAn estimate of the amount of wood processed by invertebrates c

from the field data and fecal production values.The Lara avara population would produce 0.6 g feces/kg wood/ yeaden sity of four larvae/kg wood, mean weight of 6 mg an d d aily egesbody wt. ForHeteroplectron californicum, fecal production approximateswood/year, based on a d ensity of 34 larvae/kg, mean weight of egestion rate of 50% body wt.Oxytrema silicula would prod uce 7.3 g fecwood, a ssuming 10 sn ails per kg wood, mean weight of 10 mg, an d dof 10% body wt.

T he resultan t estimate of fecal production for the three major woinvertebrates would be 17.5 g/kg of wood per year. Without snails,produced would be 10.2 g/kg of wood per year. Thus, a conservatthe direct role of invertebrates in the processing of wood would be bper year. The estimate is conservative because we have not includewood which went into xylophagous invertebrate production and resother associated in vertebrates.

•

T he time required for wood d isappearan ce in fresh water has rmated. O ne m easurement, by H odkinson (1975), predicted a 57-7.5-cm diam logs of Populus balsamifera found in a cold water beaver pond inA lberta, C an ad a. T his is considerably longer than estimates by Gilfor fine twigs less than 1.0 cm in d iam (B . Buckley, pers. comm.). Mtion of these twigs, assuming an RQ of 1 and 50% carbon compositionat 18% per year or a 5.5-year disappearance time. Wood inha bitedinvertebrates is usually in excess of 2.0 cm in diam. T hus, disappeadebris would be between the estimate of fine twigs and that of logs inpond , perhaps 10-20 years. With an estimate of 10-20 years, 1% cyear by invertebrates looks insignificant on an an nua l basis. Howeverole of invertebrates could be as much as 10-20%, considering the fulcycle.

T his approximation of wood d egrad ation by invertebrates is in as that found for litter in other terrestrial a nd aqua tic env ironm ent

8/15/2019 Anderson 1978

20/21

897 8 AND ERSON E T AL.: WOOD PRO CESSINGIN STREAMSCummins ( 1 9 7 4 ) estimated 21% invertebrate leaf processing in a Michigan stream,and our own estimate (Sedell e t a l . 1 9 7 5 ) i s 1 5 in Cascade streams. In terrestrialenvironments, Crossley and Witkamp ( 1 9 6 4 ) estimate annual litter flux through soilarthropods at 15 , and Gist and Crossley ( 1 9 7 5 ) at 10-20%.

The indirect role of feeding on senescent microflora, exposing more surface areaand spreading fungal spores, would enhance the role that these invertebrates play inthe processing of wood.

Acknowledgments.-We appreciate the field assistance of Kaye and Dan Himsworth, DaleMcCullough, Barry Frost and, especially, Cary Kerst. Robert Naiman and Stanley Gregoryprovided the data and interpretation of quantity of periphyton on wood. We thank R. W.H end erson for photographic assistance. T his man uscript would never have reached completionwithout the expert typing and patience of Alma Rogers, Paula Pitts and Larky Leeman. Dr.K W Cummins kin dly reviewed the man uscript and provided man y constructive suggestions.

T he work reported in this article was supported by N ational Science Founda tion G rant N o.GB 36810X an d N o. BM S7602656 to the Coniferous Forest Biome, Ecosystem A na lysis Studies,U. S. International Biological Program. This is contribution No. 277 from the ConiferousForest Biome, and Technical Paper No. 4402 of the Oregon Agricultural Experiment Station.

LITERATURE CITEDANDERSON, N. H. 1976. The distribution and biology of the Oregon Trichoptera.Oreg. Agric.

Exp. Stn. Tech. Bull. No. 134. 152 p.BE N EM A N N , J. R. 1973. N itrogen fixation in termites.Science, 181:164 -165.BOLING,R. H. , JR. , E. D . GO O D MAN , J. 0. ZIMMER , K. W. CUMMINS , S. R. RE ICE, R. C.

PET ER SE N AN D J. A. VA N SICKLE . 1974. Toward a model of detritus processing in awoodland stream. Ecology, 56:141 -151.

BREZNAK, J. A., W. J. BR ILL , J. W. MER TIN S A N D H . J. CO PPEL . 1973. Nitrogen fixation intermites. Nature, 244: 577-580.

BR O WN , J. K. 1971. A plana r intersect method for sampling fuel volume and surface area.For.Sci., 17:96 - 102.

BRUSVEN, M. A. 1970. Drift periodicity of some riffle beetles (Coleoptera:Elmidae). J. Kans.Entomol. Soc., 43:364 -371.

C A R P E N T E R ,E . J. AN D J. L. CU LL IN E Y. 1975. N itrogen fixation in marine shipworms.Science,187:551 -552.

CLAFLIN, T. 0. 1968. Reservoir aufwuchs on inundated trees. Trans. Am. Microsc. Soc., 87:97-104.

COFFMAN, W. P. 1978. Chironomidae, Chap. 22, p. 395-434. In: R. W. 'Merritt and K. W.Cummins (eds.). An introduction to the aquatic insects of North America. Kendall-H unt, New York.

CR O MA CK, K. AN D C. D . MO N K. 1975. Litter production, decomposition, and nutrient cyclingin a mixed hardwood watershed and a white pine watershed, p. 609-624. In F. G.

Howell, J. B. Gentry and M. H. Smith (eds.). Mineral cycling in southeastern ecosys-tems. Conf-740513. Natl. Tech. Inf. Service, Springfield, Va.CR O SS LE Y, D . A., JR . 1970. Roles of microflora and fauna in soil systems, p. 30-35.In: Pesti-

cides in the soil: Ecology, degradation and movement. Int. Symp. on Pesticides in theSoil, Feb. 25-27, 1970. Mich. State Univ. Press, East Lansing. AND M. WITKAMP. 1964. Effects of pesticides on biota and breakdown of forest litter.VIII Int. Congr. Soil Sci. (Bucharest), 3:87 -91.CUMMINS, K. W., R. C. PETE RSE N, F.0. HO WARD , J. C. WUYCHE CK AND V. I. HOL T. 1973.The utilization of leaf litter by stream detritivores. Ecology, 54:336 -345.FRENCH, J. R. 1975. The role of termite hindgut bacteria in wood decomposition. Mater. Org (Berl.), 10:1 -13.

F R O E H L I C H ,H. A. 1973. Natural and man-caused slash in headwater streams. Loggers Nandi,.33. 8 p.

GIST , C. S. AN D D . A. CR O SS LE Y, JR. 1975. A model of mineral-element cycling for an inverte-brate food web in a southeastern hardwood forest litter community, p. 84-86. In: F. G.Howell, J. B. Gentry and M. H. Smith (eds.). Mineral cycling in southeastern ecosys-tems. Conf-740513. Natl. Tech. Inf. Service, Springfield, Va.

41s :1k

N*. , 4. * /r

a *,

*

A,*

-*,* 4k\

11 *, ) *Ok • 14, k . lo k \44N. ,

8/15/2019 Anderson 1978

21/21

THE AMERICAN MIDLAND NATURALIST 100 1)HALL, J. D. AND R. L. LANTZ. 1969. Effects of logging on the habitat of coho salmon and

cutthroat trout in coastal streams, p. 355-376. In: T. G. Northcote (ed.). Symposiumon salmon and trout in streams. Inst. Fish., Univ. British Columbia, Vancouver, BritishColumbia, Canad a.

HINO, S. AND P. W. WILSON. 1958. Nitrogen fixation by a facultative bacilluJ. Bacteriol.,75:403-408.

HODKINSON, I. D. 1975. Dry weight loss and chemical changes in vascular plant litter of ter-restrial origin, occurring in a beaver pond ecosystem. J. Ecol., 63:131-142.

HYNES, H. B. N. 1970. The ecology of running waters. University of Toronto Press, Toronto,Ont. 555 p.

LAMMEL, R. F. 1972. Natural debris and logging residue within the stream environment. M.S.Thesis, Oregon State University, Corvallis. 49 p.

LYF O RD , J. H., JR. A N D S. V. GRE GO RY. 1975. The dynamics and structure ofmunities in three Cascade Mountain streams. Verh. Int. Ver. Limnol., 19: 1610-1616.

MACKAY, R. AND J. KALFF. 1969. Seasonal variation in standing crop and spinsect communities in a small Quebec stream. Ecology, 50:101-109.

MANNESMANN, R. 1972. Comparison of twenty-one commercial wood species from NorthAmerica in relation to feeding rates of the Formosan termite, Goptotermes formosanusShiraki.

Mater. Org . (Berl.), 8: 107-120.

McLAcm..A.N, A. J. 1970. Submerged trees as a substrate for benthic fauna in the recentlycreated Lake Kariba (Central Africa). J. Appl. Ecol., 7:253-266.

MECOM, J. 0. 1972. Feeding habits of Trichoptera in a mountain stream. Oikos, 23:401-407.NILSEN, H. C. A N D R. W . LA RIM O RE . 1973. Establishment of invertebrate comm

substrates in the Kaskaskia River, Illinois. Ecology, 54:366-374.PETERSEN, R. C. AND K. W. CUMMINS. 1974. Leaf processing in a woodland sFresh

water Biol., 4:343-368.ROGERS, J. S. AND G. W. BYERS. 1956. The ecological distribution, life history, andstages of Lipsothrix sylvia (Diptera:Tipulidae). Occas. Pap. Mus. Zool. Univ. Mich.No. 572. 14 p.

RO TH ACH ER , J. , C. T. D YRN ESS AN D R. L. FR ED RIKSEN . 1967. Hydrologicteristics of three small watersheds in the Oregon Cascades. U.S. For. Serv. Res. NotePNW. 14 p.

SAVORY, J. G. 1954. Damage to wood caused by microorganisms. J. Appl. Bacteriol., 17:213-218.

SEDELL, J. R., F. J. TRISKA, J. D. HALL, N. H. ANDERSON AND J. H. LYFORD. 1974. Sourcean d fates of organic inputs in coniferous forest streams, p. 5 7-69.In: R. H. Waring andR. L. Edmonds (eds.). Integrated research in the coniferous forest biome. ConiferousFor. Biome Bull. No. 5. University of Washington, Seattle.

ANDN S. TRISKA. 1975. The processing of conifer and hardwood leaves intwo coniferous forest streams. I. Weight loss and associated invertebrates. Verh. Int.Ver. Limnol., 19:1617-1627.

SWANSON,F. J., G. W. LIENKAEMPER AN DJ. R SEDELL. 1976. The history and physical

effects of large organic debris in western Oregon streams. U.S. For. Serv. Gen. Tech.Rep. PNW-56. 15 p.WA GN E R , C. E . VA N . 1968. The line intersect method in forest fuel sampling.For. Sci., 14:

20-26.WARREN, C. E., J. H. WALES, G. E. DAVIS AND P. DOUDOROFF. 1964. Trout produ

experimen tal stream enriched w ith sucrose.J. Wild. Manage., 28:617-660.W I N T E R B O U R N ,M. J. 1971. The life histories and trophic relationships of the Trichoptera of

Marion Lake, British Columbia. Can. J. Zool., 49:623-635.

SUBMITTED 8 N OVEMBER 1976

ACCEPTED 21 MARCH 1977

82