Journal of Paleolimnology20: 277–293, 1998. 277c 1998Kluwer Academic Publishers. Printed in the Netherlands.

Postglacial changes in chironomid communities and inferred climate neartreeline at Mount Stoyoma, Cascade Mountains, southwestern BritishColumbia, Canada

Michael J. Smith1, Marlow G. Pellatt1, Ian R. Walker1;2 & Rolf W. Mathewes1*1Dept. of Biological Sciences, Simon Fraser University, Burnaby, British Columbia, Canada V5A 1S6 (e-mail:mpellatt@sfu.ca; r-mathewes@sfu.ca) *Author to whom reprint requests should be sent2Dept. of Biology, Okanagan University College, 3333 College Way, Kelowna, British Columbia, Canada V1V1V7 (e-mail: iwalker@okcins.okanagan.bc.ca)

Received 22 July 1997; accepted 9 November 1997

Key words:chironomid, Holocene, paleoclimate, paleolimnology, treeline, lake sediment, British Columbia

Abstract

Analysis of the distributions of chironomid (midge) and other dipteran subfossils from two high elevation lakesediment cores in the Cascade Mountains reveals changes in midge communities and inferred climate since thelate-glacial. Cabin Lake and 3M Pond are located near treeline in the subalpine Engelmann Spruce/Subalpine Firbiogeoclimatic zone of British Columbia. In Cabin Lake, chironomid head capsule assemblages depict a typicallate-glacial community, and three distinct Holocene communities. In Cabin Lake, the late-glacial communityis composed of cold-stenothermous taxa dominated byStictochironomus, Mesocricotopus, Heterotrissocladius,Parakiefferiella nigra, ProtanypusandParacladius, whereas warm water midges are absent or rare, indicating coldconditions. A late-glacial chironomid community was not found in 3M Pond. In both lakes the early Holocene isdominated by a diverse warm-adapted assemblage,corresponding to the warm climatic conditions of the xerothermicperiod. Cabin Lake’s mid-Holocene zone records a decrease in relative abundance of the warm water types andis accompanied by an increase in cold-stenotherms. At 3M Pond this period shows a dramatic loss in diversity ofwarm-adapted taxa, as the temperate genusDicrotendipesdominates. This zone corresponds to Hebda’s (1995)mesothermic period. Further cooling in the late Holocene (to modern conditions) is inferred from continuedreduction of warm water midges and persistence (at Cabin Lake) or appearance (at 3M Pond) of a cold-stenothermalcommunity. This late Holocene cooling is similar in timing to Neoglacial advances in the Coast, Cascade, andRocky Mountains of southern British Columbia. Similarities in the timing of chironomid and vegetation communitychanges at these high elevation sites, along with the more rapid response time of the Chironomidae, support thesensitivity of midges to postglacial climatic change at high elevation sites.

Introduction

Instrumental records of climatic variation are a relative-ly recent development in human history and so provideinadequate perspectives on climatic variation over along time scale (Bradley, 1985). Paleoclimatology isthe study of climate prior to the period of instrumen-tal measurements, primarily using various proxy data,in the form of climate-dependent natural phenomena(Bradley, 1985). It is possible to synthesize differentproxy data into a comprehensive picture of former cli-

matic fluctuations, thus increasing the possibilities ofidentifying their causes and mechanisms.

This study focuses on changes in climate since thelast deglaciation in southern British Columbia, Cana-da. Most paleoclimatic studies in this region haveused paleobotanical information to infer past climatechange, but these studies show some discrepancieswith respect to the timing of climatic changes (Heb-da, 1995). However, chironomid fossils can also pro-vide valuable paleoclimatic proxy data (e.g., Levesqueet al., 1993, 1996; Walker et al., 1991a, 1991b), and

278

very pronounced differences have been noted in chi-ronomid faunas across altitudinal and latitudinal gradi-ents spanning arctic and alpine treelines (Lotter et al.,in press; Olander et al., 1997; Walker & Mathewes,1989; Walker et al., 1991a; Walker & MacDonald1995). Thus, chironomid paleoecological studies con-ducted near present treeline are likely to provide sen-sitive records of Holocene climatic change (Walker &MacDonald, 1995).

In this paleolimnological study, we use fossil chi-ronomid larval remains in the sediments of two highelevation lakes to infer changes in summer water tem-peratures since late-glacial time. Our study sites lievery near alpine treeline, which is known to be verysensitive to the effects of climate on plant growth anddistribution (Arno & Hammerly, 1984; Luckman &Kearney, 1986). By conducting our work in conjunc-tion with studies of pollen and plant macrofossils atthe same sites (Pellatt, 1996), two independent linesof proxy information are now available for late-glacialand Holocene climatic change at these sites.

Study sites

Located in the Kamloops Forest Region of the south-western interior of British Columbia, Mount Stoyoma(2283 m asl; 121�130W; 49�590N) is at the northernlimit of the Cascade Mountains (Figure 1). The Cas-cades are primarily composed of strongly folded Pale-ozoic and Mesozoic sedimentary and volcanic rocks(Holland, 1976).

The sites lie in the Engelmann Spruce-SubalpineFir biogeoclimatic zone (ESSF), which is the high-est forested zone in the southern and central interiorof British Columbia (Hebda, 1995). More specifical-ly, the study sites are within the dry, cold EngelmannSpruce - Subalpine Fir subzone variant 2 (ESSFdc2)(Lloyd et al., 1990), near treeline and just below thealpine tundra biogeoclimatic zone. Characteristic treesand shrubs include Engelmann spruce (Picea engel-mannii), subalpine fir (Abies lasiocarpa), whitebarkpine (Pinus albicaulis), white-flowered rhododendron(Rhododendronalbiflorum), and black mountain huck-leberry (Vaccinium membranaceum) (Pellatt, 1996),producing continuous forest at lower and middle ele-vations in the zone, and subalpine parkland at higherelevations (Lloyd et al., 1990).

The continental climate is relatively cold, moist,and snowy, with a short growing season, and long,cold winters. Mean annual temperature is�2 to 2�C,

with 5 to 7 months below 0�C and less than twomonths above 10�C (Coupe, 1983; Pojar et al., 1991).Precipitation ranges from 400 mm in the drier portionsto 2200 mm in the wetter areas. As much as 50 to 70%of the precipitation falls as snow.

The two lake basins selected for paleoecologicalanalysis were Cabin Lake (1850 m asl) and 3M Pond(1950 m asl). Within the ESSFdc2, Cabin Lake (Fig-ure 2) lies within the dry southern Engelmann Spruce– Subalpine Fir Parkland and is surrounded by a welldeveloped forest of Engelmann spruce and subalpinefir on the south, east, and west sides. The north slopeforest was burned relatively recently, having abundantcharred snags, conifer seedlings, and open meadowvegetation.

Cabin Lake has an area of approximately 4 ha, anda maximum water depth measured at 4.2 m. A smallinlet stream is located at the north end which is activeduring the warmest months as snow melts at higherelevations, and an outlet stream is present at the southend, perhaps also only active when meltwater raiseslake level.

Dominant trees and shrubs surrounding CabinLake include Engelmann spruce, subalpine fir, white-flowered rhododendron, partridgefoot (Luetkea pecti-nata), black mountain huckleberry, grouseberry (Vac-cinium scoparium), pink mountain heather (Phyl-lodoce empetriformis), and mountain heather (Cas-siope mertensiana) (Pellatt, 1996; Lyons, 1965).

3M Pond (Figure 3) lies within the transitionbetween ESSF parkland and the Dry Southern AlpineTundra (ATd) (Mitchell & Green, 1981). The pondis approximately 0.75 ha in area and has a maximumwater depth of 1 m. Common trees and shrubs sur-rounding the pond are subalpine fir, Engelmannspruce,whitebark pine, common juniper (Juniperus commu-nis), white mountain heather, pink mountain heather,and partridgefoot, with a more prominent herb compo-nent than observed around Cabin Lake (Pellatt, 1996).

Methods

Field and laboratory

Cabin Lake was sampled in July, 1994 using a modi-fied Livingstone piston corer with a core tube diameterof 5 cm. A 399 cm sediment core was obtained fromthe deepest part of the lake. The five drives comprisingthe entire core length were extruded in the field andwrapped in cellophane and aluminium foil, and then

279

Figure 1. Map of British Columbia showing the location of Mount Stoyoma in the Northern Cascade Mountains. Inset shows location of CabinLake and 3M Pond.

280

Figure 2. Photograph of Cabin Lake.

Figure 3. Photograph of 3M Pond.

transported to Simon Fraser University (SFU) wherethey were stored at approximately 4�C for later analy-sis.

The Cabin Lake sediments were subsampled mostcommonly every 15 cm, with higher resolution sam-

pling in regions of expected faunal change (i.e., late-glacial basal clay and basal clay/gyttja interface). Sub-samples consisted of 0.5 ml or 1 ml of sediment, but upto 10 ml of sediment were necessary in some intervalsto obtain sufficient numbers of chironomid head cap-

281

Table 1. AMS radiocarbon dates for 3M Pond and Cabin Lake

Sample identification Sample description Lab # Age (14C yr BP)

3M Pond (65–67 cm) conifer needles TO-5329 10 000�320

3M94-B-21 (Isotrace)

3M Pond (24–26 cm) conifer needles TO-5330 3530�60

3M94-A-25 (Isotrace)

Cabin Lake (306 cm) carbonized wood TO-5205 8910�120

CLJL 94-5-12 (Isotrace)

Cabin Lake pollen 29826 10 090�70

CL-324 cm (CAMS)

Cabin Lake leaf fragments 29829 9860�60

CL-368-370 cm (CAMS)

(Errors presented as�1 s).

sules for analysis. Counts per sample ranged from 30to 178.5 head capsules, with a mean of 73.2. All sub-samples were stored in glass vials in 2 ml of distilledwater and kept refrigerated at 4�C until analyzed.

Coring of 3M Pond used the methods outlinedabove in August, 1994. A 73 cm sediment core wasremoved from the deepest region of the pond and trans-ported to SFU.

The 3M Pond core was subsampled at 2 cm or4 cm intervals in the top half of the core (0–35 cm),and every 8 cm in the bottom half (35–67 cm). Again,0.5 ml or 1 ml (but up to 2 ml) subsamples were placedin 2 ml of distilled water in glass vials and stored atapproximately 4�C for future analysis. Number ofhead capsules per sample ranged from 36 to 460.5,with a mean of 180.2.

Isolation of chironomid head capsules,Chaoborusmandibles, and ceratopogonid head capsules general-ly followed the procedures outlined by Walker (1987).Sediment samples were deflocculated in 8% KOH atapproximately 60�C for 60 min, followed by a thor-ough distilled water rinse on a 95�m Nitex mesh. Theresidue retained on the sieve was back-washed into abeaker using distilled water, in which it was storeduntil sorted.

The residue was sorted in a Bogorov counting trayunder 25–50 X magnification with a Wild M5 dis-secting microscope. Head capsules were transferred todrops of water on a coverslip under 12 X magnificationusing #4 or #5 forceps. Coverslips containing approxi-mately 30 head capsules each were dried and mounted

onto glass slides using Permount or Entellan mountingmedium for later identification.

Remains were identified under 100–400 X magni-fication using various Zeiss compound microscopes.Identifications were based on descriptions given inkeys by Walker (1988), Oliver & Roussel (1983) andWiederholm (1983). Whole head capsules, and frag-ments containing greater than half of the mentum, werecounted as one head capsule. Fragments that wereexactly half of a head capsule were counted as onehalf, and fragments that included less than half ofthe mentum were not counted. Most identificationswere made at the generic level, although a few speciesidentifications were possible. Broader taxonomic cat-egories were necessary where the genus could notbe determined (e.g., subtribe Tanytarsina,Cricoto-pus/Orthocladius, Corynoneura/Thienemanniella).

Data analysis

Data were compiled on a spreadsheet using TILIA ver-sion 2.0 (Grimm, 1993), and chironomid percentagediagrams were producedusing TILIA-GRAPH version1.25 (Grimm, 1993). A constrained sum-of-squarescluster analysis (CONISS) (Grimm, 1987) for percent-age data was done to examine major changes in chi-ronomid communities throughout the stratigraphy.

282

Figure 4. Chironomid percentage diagram for Cabin Lake, with 10X exaggeration curves (stippled) to highlight the abundance of rare taxa.Data are archived with NOAA as TILIA files.

Chronology

Ash layersTephra derived from Mount Mazama (6730 yr BP)(Hallett et al., 1997) is present in both Cabin Lake(246–253 cm) and 3M Pond (36–42 cm) sedimentprofiles. Cabin Lake also contains the Bridge RiverTephra (ca. 2410 yr BP) (Clague et al., 1995) at 89.5–90.5 cm. Identifications of tephra using microprobeanalysis were made by Dr Gerald Osborn and GlenDePaoli (University of Calgary).

Radiometric datingAMS radiocarbon dating (by IsoTrace RadiocarbonLaboratory, Toronto, and Center for Accelerator Mass

Spectrometry, Lawrence-Livermore National Labora-tory) was used to determine the ages of organics at twopoints in 3M Pond’s sediment profile, and three pointsfor Cabin Lake, and all dates given in this study are inuncalibrated radiocarbon years before present (14C yrBP), datum 1950. Pollen dates were obtained using themethod of Brown et al. (1989). Table 1 shows sedimentdepths, types, and ages of samples from Cabin Lakeand 3M Pond. Regression analysis (Grimm, 1993) wasused to interpolate radiocarbon ages at key intervals ofchironomid community change.

283

Figure 4. Continued.

Results

Stratigraphy and chronology

Cabin LakeThe basal sediments (399–324 cm) of Cabin Lake con-sist of a dark grey clay with two distinct bands (369–364 cm, 357–355 cm) of very dark grey mixed organ-ic/inorganic material which are graded coarse to fine.A relatively uniform dark brown to black gyttja wasdeposited above 324 cm. The remainder of the sedi-ment profile consists of dark gyttja, with the exceptionof two tephras at 252.5–247 cm (identified as Maza-ma ash) and 90.5–89.5 cm (identified as Bridge Rivertephra).

A sample of carbonized wood at 306 cm was dat-ed at 8910�120 yr BP (Table 1). Dates of pollen at324 cm and leaf fragments at 368–370 cm (both with-in the clay) were 10 090�70 and 9860�60 yr BP,respectively. Therefore, it was assumed that the claywas deposited rapidly at the late-Pleistocene/Holoceneboundary at ca. 10 000 yr BP.

3M PondThe core taken from 3M Pond contains a uniformlygrey basal clay from 73–65.5 cm. Above 65.5 cm, var-ious shades of gyttja, ranging from black to olive greyto the green surface sediments, make up the remainderof the core. Mazama ash (6730 yr BP) was also presentin the stratigraphy between 46–42 cm. Bridge Rivertephra was not evident in 3M Pond sediments. Ash

284

Figure 5. Chironomid percentage diagram for 3M Pond, with 10X exaggeration curves (stippled) to highlight the abundance of rare taxa. Dataare archived with NOAA as TILIA files.

is often washed in by surrounding slopes and inlets,which are insignificant around 3M Pond, and a verythin ash layer deposited in the lake basin may havebeen disturbed by bioturbation.

In addition to the dated tephra, two AMS radio-carbon dates were obtained. Conifer needles foundat the clay/gyttja interface (66 cm) produced a dateof 10 000� 320 yr BP, corresponding with the late-Pleistocene/Holocene boundary, as expected. Similarneedles were isolated from 26–24 cm (near major chi-ronomid zone boundary), giving a date of 3530� 60 yrBP (Table 1).

Chironomids

Figures 4 & 5 illustrate head capsule counts for eachtaxon, as a percentage of the total number of identifi-able chironomid head capsules in each sampling inter-val. The chironomid taxa, along withChaoborusandCeratopogonidae, are separated into typical warm andcold water assemblages according to current informa-tion regarding their distributions (Walker, 1990; Walk-er & MacDonald, 1995; Walker & Mathewes, 1989;Walker et al., 1997). The remaining uncategorizedtaxa consist of ‘rheophilous’ (flowing water) groups(i.e., Brillia/Euryhapsis, Doithrix/Pseudorthocladius,Corynoneura/Thienemanniella,Rheocricotopus, Euki-efferiella & Tvetenia, Parametriocnemus, Smit-

285

Figure 5. Continued.

tia/Pseudosmittia, Synorthocladius), eurythermicgroups (i.e.,Psectrocladius, Procladius), and broadtaxonomic categories (i.e.,subtribe Tanytarsina,Crico-topus/Orthocladius), all of which are of relatively littlevalue in assessing paleotemperature regimes.

Cabin LakeThe chironomid fossil record of Cabin Lake has beendivided into four zones, determined by a constrainedsum-of-squares cluster analysis (Grimm, 1987) (Fig-ure 4).

Zone CC-1 (399–330 cm;>10 000 yr BP)The late-glacial assemblage primarily consists of thewidespread Tanytarsina group (up to 54%) and typicaloligotrophic, cold-stenotherms (Paracladius, Paraki-

efferiella nigra, Mesocricotopus, Stictochironomus,Protanypus, andHeterotrissocladius). Interesting fea-tures of this zone are the brief peaks in the relativeabundances of some typical warm-adapted chirono-mids (Chironomus, Stempellinella& Zavrelia, Paraki-efferiella cf. bathophila, Dicrotendipes, Stempellina)and theDasyhelatype of ceratopogonid. They occurat the time of increased deposition of organic sedimentprior to the Holocene.

The predatoryProcladiusalso makes up a signif-icant proportion of the late-glacial assemblage (up to33%), and remains relatively abundant throughout theHolocene.

286

Zone CC-2 (332–265 cm; ca. 10 000 to 7220 yr BP)The beginning of the Holocene is characterized by asudden disappearance of most of the cold-stenotherms,with only Heterotrissocladiuspersisting as a verysmall proportion of the faunal assemblage. In con-junction with this trend, significant increases in thewarm-adapted chironomids occur.Chironomus, Stem-pellinella& Zavrelia, Pagastiella, andMicrotendipes,along with variousChaoborusspecies, rapidly becomemajor faunal elements in the early Holocene. TheeurythermicTanytarsina andProcladiusapproach theirlowest levels in this zone, which may be a result of low-er production in these groups or simply because of agreater prevalence of other taxa. Unfortunately, sed-imentation rates are not available and so influx ratescould not be determined.

This zone also shows an increase in proportions ofthe rheophilousCorynoneura/Thienemanniellagroup,and the eurythermicPsectrocladiusandCorynocera.

Zone CC-3 (265–112 cm; 7220 to 3210 yr BP)Just prior to Mazama ash deposition, a major shift infauna occurs. Although this mid-Holocene zone stillsupports a large warm-adapted group of chironomids,a notable reduction inStempellinella& Zavrelia andPagastiellais seen, andSergentiareaches significantproportions (up to 27%) of the community composi-tion. Heterotrissocladiusalso reaches levels compara-ble to its late-glacial contribution.

Chironomusremains a dominant genus through thiszone (up to 27%), withMicrotendipesconsistentlymaking up approximately 5% of the community. Oth-er important warm-water taxa comprising this zone areDicrotendipesandParakiefferiellacf. bathophila, withChaoborusmandibles reaching their greatest propor-tion.

Tanytarsina (37%) andProcladius (34%) againattain significant relative abundances in the mid-Holocene, and various rheophilous taxa continue tobe consistently represented in small proportions.

Zone CC-4 (112–0 cm; 3210 yr BP to present)The most recent late Holocene sediments reveal a dra-matic decrease in the relative contribution of warmwater taxa as a group, with many genera becom-ing locally extinct and others reaching their minimalHolocene values.Sergentiaremains relatively abun-dant, withHeterotrissocladiusretaining its typical late-glacial abundance at the beginning and end of this inter-val. Possibly significant is the reappearance ofPara-

cladius, a strongly cold-stenothermous taxon (Walkeret al., 1997), in the late Holocene fossil assemblage.

Procladius remains a dominant faunal element,whereas Tanytarsina reaches its maximum extent(61%), surpassing its contribution to the late-glacialassemblage. Rheophilous taxa continue to be deposit-ed into the lake, although a slightly less diverse groupis evident in the late Holocene.

3M PondChanges in chironomid communities in 3M Pond canalso be divided into three Holocene zones (Figure 5),as for Cabin Lake. Rarity of head capsules in the clayprevented the extraction of information regarding thelate-glacial at 3M Pond.

Zone 3MC-1 (66–44 cm; 10 000 to 6730 yr BP)The early Holocene was dominated by Tanytarsi-na, making up approximately 60% of the faunalassemblage. The eurythermicPsectrocladius(7–16%)and Procladius (up to 12%) also significantly con-tribute to this assemblage. A distinct and diversewarm water assemblage is clearly evident, com-posed primarily of the tribe Pentaneurini,Chirono-mus, Cladopelma,Polypedilum, Microtendipes, Dicro-tendipes, Chaoborus, and other minor components.Not surprisingly, a lack of cold-stenothermic chi-ronomids is apparent, with only threeHeterotrisso-cladius head capsules being found. A very minorallochthonous contribution is possible as indicated bythe presence ofCricotopus/Orthocladius, Corynoneu-ra/Thienemanniella, Limnophyes, and Rheocricoto-pus.

Zone 3MC-2 (44–27.5 cm; 6730 to 3950 yr BP)Following Mazama ash deposition, a major shift in thefaunal assemblage occurred, with most of the warm-adapted taxa becoming locally extinct or reduced to avery minor component of the community. An excep-tion to this trend is the rapid increase in the relativeabundance ofDicrotendipes(up to 80%), also a typ-ical temperate genus. A significant reduction in therelative abundances of Tanytarsina andPsectrocladiusmay be the result of decreased production or an artefactof the dominance ofDicrotendipes. Again, influx dataare lacking, and so the cause of this change remainsuncertain. The contributions ofProcladius, Cricoto-pus/Orthocladius, andCorynoneura/Thienemanniellashow little change from the early Holocene.

287

Zone 3MC-3 (27.5–0 cm; 3950 yr BP to present)Another major zone boundary is evident at 27.5 cm,whenDicrotendipesrapidly declines in relative abun-dance to near extinction. The only other warm-adaptedchironomid,Microtendipes, persists for a short timeas a very minor faunal element. Offsetting the disap-pearance ofDicrotendipesis the littoralPsectrocladius(up to 57%), which dominates the late Holocene alongwith Tanytarsina (32–72%).

The increased abundance of cold-stenotherms (Het-erotrissocladius, Stictochironomus) in this zone is sig-nificant, as it represents their first appearance as a dis-tinct community in the Holocene sediments.

Procladius, Cricotopus/Orthocladius, andCory-noneura/Thienemanniellaremain relatively unchangeduntil the most recent sediments, and there again existsa minor rheophilous contribution, as in the earlyHolocene. The appearance ofDerotanypusin the lateHolocene freshwater assemblage is surprising, as thisgenus is common in British Columbian saline lakes.

Discussion

Changes in chironomid assemblages through post-glacial time could be the result of many different envi-ronmental influences, but at Cabin Lake and 3M Pondon Mount Stoyoma, these community shifts can beexplained by direct or indirect climatic effects, whichclosely correspond to other Holocene paleoclimatereconstructions in southwestern British Columbia.

Cabin LakeZone CC-1The late-glacial midge assemblage clearly contains asignificant cold-stenothermous community, includingParacladius, Parakiefferiella nigra, Mesocricotopus,Stictochironomus, Protanypus, and Heterotrissocla-dius(Figure 4). This diverse cold-adapted assemblage,along with the rarity of typical temperate, warm waterchironomids, provides evidence of a cold, late-glacialclimate at this high elevation site. Alpine tundra con-ditions during late-glacial time are also indicated bythe presence of subalpine/alpine herb pollen in thesesediments (Pellatt, 1996).

Zone CC-2The beginning of the Holocene is accompanied by adramatic reduction of cold-stenotherms, asHeterotris-socladiusbecomes very rare, and other typical late-

glacial taxa become locally extinct. The rapid diver-sification and dominance of warm water taxa in theearly Holocene suggests a shift from late-glacial coldconditions to a warm climate, which persisted to ca.7220 yr BP (Figure 4). Warmer water temperatureswould inhibit cold-stenothermous taxa from complet-ing their life cycles (Oliver, 1968, 1971; Danks & Oliv-er, 1972a, 1972b), whereas the warm water taxa wouldencounter temperatures sufficiently high for comple-tion of their life cycles.

Although it may appear that the decline of olig-otrophic taxa and dominance of the eutrophic indicatorChironomussuggests a eutrophication process in theearly Holocene, it should be recalled that Cabin Lakewas shallow, probably with a maximum depth lessthan 7.5 m at that time, and so the lake was likely verywell mixed in the summer. Even if the lake stratified,the profundal region would have been small, and thevast majority of the head capsules that were depositedwere probably derived from the much more extensivelittoral habitats. Thus, it would be unwise to applythe lake typology system of Sæther (1979) to explainthe chironomid community shift at the beginning ofthe Holocene. Kansanen & Aho (1981) and Hofmann(1978) explain that in unstratified lakes bottom tem-peratures can be high, and so the cold-stenothermaldeep-water species, which are usually important com-ponents of the oligotrophic fauna, may be absent inshallow, oligotrophic lakes, not because of eutrophyand its associated summer bottom oxygen depletion(Hofmann, 1986, 1988; Walker, 1987), but simply dueto high temperatures.

Hofmann (1983, 1986) notes that during late-glacial cold conditions, the cold-stenothermic commu-nity common in oligotrophic waters may have occurredin the littoral zone, as is observed in arctic and sub-arctic lakes, and so would not reflect trophic state oroxygen conditions in deeper waters.

A community dominated byChironomususuallysuggests eutrophic conditions, but a study of chirono-mids in Lake Mallasvesi, Finland, challenges this gen-eralization (Kansanen et al., 1984). Lake Mallasvesi isoligotrophic to mesotrophic and irregularly stratified,with high hypolimnetic temperatures in both summerand winter. Thus, rather than indicating trophic state,the Chironomuscommunity reflects climatic condi-tions, either directly through temperature effects on thelarvae, or indirectly, as high temperatures would causeincreased oxygen consumption and lead to decreasedoxygen concentration in deep waters (Kansanen et al.,1984).

288

Pellatt’s (1996) vegetation analysis of Cabin Lakepostglacial sediments reveals total pollen concentra-tions at their highest in the early Holocene. High extra-local or regional transport of diploxylonpine andAlnusviridis pollen suggest that open/dry conditions existedaround Cabin Lake in the early Holocene. Relativelyhigh values ofPicea, haploxylon pine,Artemisia, andPoaceae pollen support this hypothesis (Pellatt, 1996).Thus, agreement of the paleoclimate reconstruction(with respect to temperature) using fossil vegetationand chironomid community composition is consis-tent with coastal British Columbia’s early Holocenexerothermicperiod (Mathewes, 1973, 1985; Mathewes& Heusser, 1981; Hebda, 1995).

Zone CC-3The chironomid zone encompassing the mid-Holocene(ca. 7220 yr BP to ca. 3210 yr BP) supports a welldefined warm water chironomid community, althoughthere is a significant decrease in the relative abundancesof Stempellinella/ZavreliaandPagastiella(Figure 4).A significant increase in relative abundance ofSer-gentia can also be noted.Sergentiais often foundin cold waters (Walker, 1990; Walker & Mathewes,1989; Walker et al.,1997), but has also been reportedin warmer waters such as those found in arctic ponds(Walker & Mathewes, 1989).

In considering temperature requirements alone,it appears that the reduction in some warm watertaxa suggests cooling following the early Holocene(Zone CC-2), but conditions still warmer than present.The prevalence ofSergentiamay be a result of win-ter anoxia (Walker & Mathewes, 1989a), possibly dueto increased precipitation in the mid-Holocene, as thisgenus thrives in waters with moderate oxygen deple-tion (Walker, 1990). Although not discernible fromthe chironomid assemblage directly, more precipita-tion would produce greater snow pack on the lake andlead to longer winter conditions, as the time requiredfor the snow to melt would be extended and anoxicconditions would persist for a longer period. A smallbut more diverse assemblage of rheophilous taxa in thiszone may suggest increased precipitation, as a greatervolume of melting snow from higher elevations couldcarry more stream taxa into Cabin Lake.

The presence of pollen types indicative of relative-ly wet conditions (i.e.,Abies, Cyperaceae, Ericales)reveal that conditions were wetter during the mid-Holocene than in the late-glacial or early Holoceneat Cabin Lake (Pellatt, 1996). In addition, high pollen

concentrations and low spruce/pine ratios indicate aclimate warmer than present, while an increase insome subalpine taxa suggests a decrease in tempera-ture with respect to the early Holocene (Pellatt, 1996).Therefore, the chironomids and vegetation at CabinLake both indicate a climate that corresponds withthe mesothermic period noted in southwestern BritishColumbia (Hebda, 1995).

Pellatt (1996) distinguishes a second pollen zone inthe more recent sediments of chironomid zone CC-3,astypical subalpine vegetation becomes established and afurther decrease in temperature is inferred (Figure 6).This continued cooling of the latter part of the mid-Holocene is not as obvious in the chironomid stratigra-phy, although the cold-stenothermHeterotrissocladiusbecomes more abundant and the relative contributionof Chironomusdeclines. As Pellatt (1996) suggests,the mid-Holocene may have been a time of contin-ued gradual cooling rather than a stable temperaturestate, and this is evident from the chironomid profile,although further zones are not as distinct. Higher reso-lution chironomid sampling may have better illustratedthis subtle trend.

Zone CC-4The late Holocene (ca. 3210 yr BP to present) broughta further reduction in the relative abundances of thewarm water taxa, as many genera became locallyextinct and others reached their lowest Holocene values(Figure 4). Conversely, the cold-stenothermousHet-erotrissocladiusessentially retained its mid-Holocenecommunity contribution, andParacladius, althoughrare, appeared again for the first time since the lateglacial.

This significant change in the chironomid assem-blage suggests further cooling in the late Holocene,andis consistent in timing to glacial advances in the Cana-dian Rocky Mountains, Coast Mountains and northernCascade Mountains (Porter & Denton, 1967; Ryder &Thomson, 1986; Luckman et al., 1993).

3M PondZone 3MC-1The most prominent feature of the early Holocene withrespect to paleoclimatic reconstruction is the diversewarm water midge community and absence of cold-stenotherms, indicating that the water of 3M Pond wasquite warm. With a maximum depth of less than 1.7 mduring this time, the shallow waters would warm veryquickly after snow and ice melt, and likely attain suf-

289

Figure 6. Comparison in timing of zone changes for chironomids and vegetation at Cabin Lake and 3M Pond. Note: Approximate dates of zoneboundaries were interpolated through regression analysis (Grimm, 1993).

ficiently high temperatures to exclude cold water taxa(Hofmann, 1983; Kansanen & Aho, 1981; Kansanen,1985; Walker & Paterson, 1983). Pollen and plantmacrofossil analyses of early Holocene sediments indi-cate relatively dry conditions, but indicators of warmtemperatures, other than high pollen concentrations,are lacking (Pellatt, 1996). Taken together, palynolog-ical and chironomid evidence suggest that the earlyHolocene was warm and dry from 10 000 yr BP untilthe time of Mazama ash deposition, coinciding withthe observed xerothermic in coastal British Columbia(Mathewes, 1985; Hebda, 1995) and the southern inte-rior of British Columbia (Mathewes & King, 1989;Hebda, 1995).

Zone 3MC-2The dramatic shift in chironomid community com-position following Mazama ash deposition involvesthe local extinction of most warm water taxa, withthe exception of the rapid and substantial increasein relative abundance of the warm-adaptedDicro-tendipes(Figure 5).This major change does not include

the appearance of cold-stenothermous taxa, but thedecreased richness and evenness (diversity), especial-ly involving warm water chironomids, may indicatecooling.

Hoffman et al. (1996) have shown that the num-ber of nearshore macroinvertebrate taxa is positivelycorrelated to water temperature and inversely relatedto elevation in lakes of Washington’s North CascadesNational Park Service Complex. Low elevation for-est lakes were found to have the highest number oftaxa and highest maximum lake temperatures, where-as alpine sites had lowest number of taxa and lowestmaximum temperatures. Walker & Mathewes (1989)indicate that a decrease in chironomid diversity withincreasing elevation in coastal British Columbia mayreflect a relationship between diversity and tempera-ture, although they warn against uncautious interpreta-tions based on assemblage diversity alone.Smol (1981)suggests great care in making paleoecological inter-pretations of diversity measures, preferring that inter-pretations be based on the ecology of specific taxa.Recall that in the case of 3M Pond, the loss of diver-

290

sity involves warm water taxa specifically. Thorp &Chesser (1983), in their study of larval midges in acooling reservoir in South Carolina, show that watertemperature was the environmental factor most strong-ly associated with changes in chironomid assemblagediversity, with a significant relationship between diver-sity and average and maximum water temperatures.More recently, Levesque et al. (1996) found that bothchironomid species richness and diversity were lowestduring the coldest parts of the late-glacial (followingdeglaciation and during the Younger Dryas) in fourNew Brunswick lakes, whereas during the warmerperiods richness and diversity were higher.

Consistent with this interpretation, Pellatt (1996)infers mid-Holocene temperatures as warmer thanpresent from plant macrofossils and high levels of localconifer pollen, but cooler than in the early Holocene.Thus, the chironomid and vegetation compositionswithin this zone correspond well with Hebda’s (1995)mesothermic period.

The reason for the dominance ofDicrotendipesoverthe other warm water taxa is not clear. The majority ofthis genus’ larvae are found on the surface of aquaticvegetation, and among vegetation on rocks, logs, orsimilar substrata (Epler, 1988). The larvae also feed ona great variety of organic substrates, from coarse to finedetrital particles, to algae and vascular plants (Coff-man, 1984). A possible explanation for the successof Dicrotendipesafter the deposition of Mazama ashmay concern habitat and food choices. The accumula-tion of volcanic ash in the shallow basin of 3M Pondcould have greatly disturbed most habitats in whichchironomid larvae were living, butDicrotendipeswasrelatively less affected than the other taxa at this time.This may have been due to the flexibility of food typesand the ability of the larvae to live on the large surfacearea of vegetation rather than directly on the sedimentsurface.

Tokeshi (1995) reports that for epiphytic speciesin one study, the time to reach 95% of the mean nat-ural density after complete denudation of a macro-phyte habitat was less than two weeks, and Davies(1976) notes that in Volta Lake, the Chironomidae, andspecificallyDicrotendipes, dominated the colonizationof new substrates within 25 days.

Although 3M Pond did not support a large macro-phyte community throughout the Holocene, Pellatt(1996) found that the number ofCarexseeds increasedat approximately 6730 yr BP, with a complementaryincrease in Cyperaceae (sedge) pollen percentages atthis time. Sedges would have been growing on the mar-

gins of 3M Pond, and it is likely that many plants werepartly submerged. If this was the case, the larvae ofDicrotendipeswould have had the necessary habitatavailable to be relatively less affected by disruption byMazama ash deposition. An expanded knowledge ofthe ecology of each chironomid species is necessaryto support this type of interpretation of communitychanges.

A further possibility is that the ash may have insec-ticidal properties for true benthic taxa, with the sharpglass shards piercing their relatively soft body cover-ing (Miller, 1987). Chironomid taxa dwelling on plantswould not be exposed to this potentially lethal substra-tum.

Zone 3MC-3The main shift in chironomid assemblages betweenthe middle and late Holocene occurred at 28 cm in thecore (ca. 3950 yr BP) (Figure 5). This change involvedthe dramatic reduction inDicrotendipesrelative abun-dance, which was replaced primarily by the euryther-mic Psectrocladius. Thus, the late Holocene assem-blage contained few warm water taxa, withDicro-tendipesandMicrotendipespersisting very rarely andsporadically. Accompanying this trend is the appear-ance of a cold-stenothermic community which includesStictochironomusand Heterotrissocladius. Althoughthe cold-stenothermsare not a dominant group in termsof the community as a whole, their presence, alongwith the disappearance of warm-adapted taxa, indi-cates cooling in the late Holocene.

The timing of this shift in the community coin-cides, as at Cabin Lake, with the timing of glacialadvances in northwestern North America (Porter &Denton, 1967; Ryder & Thomson, 1986). As the cli-mate cooled at 3M Pond, water temperatures wouldhave decreased, causing the exclusion of chirono-mid larvae with high temperature optima and toler-ances, attaining cool enough summer temperatures tosupport cold-stenothermal taxa. This correlates wellwith the cool, modern subalpine conditions indicat-ed by decreased needle abundance and diversity, anddecreased local pollen productivity at the site afterca. 3530 yr BP (Pellatt, 1996).

A clear correlation in the timing of the changes inchironomid assemblages and vegetation at both CabinLake and 3M Pond since the late-glacial is apparent(Figure 6).

The late-glacial cold-stenothermal chironomidcommunity found in the Cabin Lake basal sediments,

291

along with the rarity of temperate taxa, indicatescold conditions before ca. 10 000 yr BP. After thistime, diverse warm water chironomid assemblagescomprise a significant proportion of the communitiesat Cabin Lake and 3M Pond, corresponding to thewarm early Holocene xerothermic period (Mathewes,1985). Palynological evidence from Cabin Lake (Pel-latt, 1996) suggests that this shift from late-glacial coldto early Holocene warm (and dry) conditions laggedbehind the midge response by approximately 700 years(ca. 9320 yr BP). Pollen and macrofossil changes at 3MPond occur at ca. 10 000 yr BP, suggesting a climate-related vegetation change at the more sensitive higherelevation site, about 700 years prior to the lower ele-vation site.

The early Holocene xerothermic period persisted toca. 7220 yr BP at Cabin Lake, at which time coolingis indicated by an increase in cold-stenotherms and areduction in temperate midge taxa. At 3M Pond, themajor decrease in relative abundance of warm-adaptedchironomids occurred slightly later, at ca. 6730 yrBP. Evidence from pollen indicates local vegetationchanges corresponding primarily to increasing mois-ture at ca. 6730 yr BP at 3M Pond and ca. 7000 yrBP at Cabin Lake (Pellatt, 1996). The midge-inferredcooling, and vegetation changes indicative of increas-ing moisture, closely correspond to the timing of themesothermic period found in coastal and southwest-ern interior British Columbia sites (Hebda, 1995), allof which appear shortly after the widespread coolingca. 7500 yr BP, evident in Greenland ice-core proxies(Alley et al., 1997).

The termination of the mesothermic appears to besomewhat time-transgressive. Earliest indications ofpossible slight cooling are found in the Cabin Lakepollen assemblage at ca. 4800 yr BP, although a moredramatic drop in temperature is evident at ca. 2410 yrBP (Pellatt, 1996). Chironomid evidence suggestsincreasing cooling at ca. 3210 yr BP at Cabin Lake,and slightly earlier at ca. 3950 yr BP at 3M Pond. Veg-etation response lagged behind chironomid response at3M Pond, where shifts in pollen at ca. 2870 yr BP andplant macrofossils at ca. 3530 yr BP indicate decreas-ing temperatures (Pellatt, 1996).

The termination of the mesothermic period at Cab-in Lake and 3M Pond coincides with Neoglacialadvances. The Tiedemann advances on the CoastMountains of British Columbia have been dated atapproximately 3345 yr BP to 1300 yr BP (Ryder &Thomson, 1986), while the Burroughs Mountain Stademoraines of the Cascade Range were built ca. 3500 yr

BP (Porter & Denton, 1967). In the Rocky Mountains,the Peyto, Robson, Yoho and Saskatchewan glaciersadvanced between ca. 3300 yr BP and 2800 yr BP(Luckman et al., 1993). Earlier indications of cool-ing have been outlined for the Rockies at ca. 5000 yrBP (Luckman et al., 1993), the Coast Mountains atca. 6000 yr BP (Garibaldi phase; Ryder & Thomson,1986), and the Southern Cascade Glacier of the north-ern Cascade Mountains at ca. 4700 yr BP (Porter &Denton, 1967).

As Porter & Denton (1967) and Ryder & Thomson(1986) point out, the transition from relatively warmconditions of the early to mid Holocene to cooler andmoister conditions was prolonged up to 2000 years,and the timing of the start of glacial advances variedregionally. Therefore, although the mesothermic peri-od in southwestern British Columbia indicates wetterand slightly cooler conditions than the early Holocenexerothermic, its termination cannot be correlated withNeoglaciation throughout this entire region.

Conclusions

In summary, this study has shown that the Chirono-midae are valuable in paleoclimatic reconstructions atsensitive high elevation sites, and that response times ofmidge communities to temperature changes are gener-ally shorter than that of vegetation, as would be expect-ed by the mobility and much shorter life cycles of chi-ronomids. Similarities in the timing of vegetation andchironomid community changes on Mount Stoyoma,and their correlation with inferred climate shifts at oth-er southwestern interior and coastal British Columbiasites clearly illustrate that shallow lakes at elevation-al treeline ecotone boundaries are very useful areasfor future paleoclimatic studies involving chironomid-inferred temperaturechanges. Analysis of surface sam-ples from a number of lakes in southwestern BritishColumbia is in progress, which will be utilized in pro-ducing a transfer function to quantitatively reconstructpostglacial temperature changes from changing chi-ronomid communities in lake sediment records in thisregion.

Acknowledgments

We wish to thank Tisha Larsson, Jenyfer Newmann,and Christine Smith for their help in the tedious taskof sediment processing and head capsule isolation.

292

Assistance in chironomid larval head capsule iden-tification was provided by Samantha Palmer, AndreJ. Levesque and Markus Heinrichs, and was very muchappreciated. This research was supported via ResearchGrants awarded by the Natural Sciences and Engineer-ing Research Council (NSERC) of Canada to IRW andRWM.

References

Alley, R. B., P. A. Mayewski, T. Sowers, M. Stuiver, K. C. Taylor& P. U. Clark, 1997. Holocene climatic instability: A prominent,widespread event 8200 yr ago.Geology25: 483–486.

Arno, S. F. & R. P. Hammerly, 1984. Timberline – Mountain andArctic Forest Frontiers. The Mountaineers, Seattle, 304 pp.

Bradley, R. S., 1985. Quaternary Paleoclimatology. Allen and UnwinInc., London, 472 pp.

Brown, T. A., D. E. Nelson, R. W. Mathewes, J. S. Vogel &J. R. Southon, 1989. Radiocarbon dating of pollen by acceler-ator mass spectrometry. Quat. Res. 32: 205–212.

Clague, J. J., S. G. Evans, V. N. Rampton & G. J. Woodsworth,1995. Improved age estimates for the White River and BridgeRiver tephras, western Canada. Can. J. Earth Sci. 32: 1172–1179.

Coffman, W. P., 1984. Chironomidae, In R. W. Merritt & K. W. Cum-mins (eds), An Introduction to the Aquatic Insects of North Amer-ica, 2nd ed. Kendall/Hunt, Dubuque: 551–652.

Danks, H. V. & D. R. Oliver, 1972a. Seasonal emergence of somehigh arctic Chironomidae (Diptera). Can. Entomol. 104: 661–686.

Danks, H. V. & D. R. Oliver, 1972b. Diel periodicities of emergenceof some high arctic Chironomidae (Diptera). Can. Entomol. 104:903–916.

Epler, J. H., 1988. Biosystematics of the genusDicrotendipesKief-fer, 1913 (Diptera: Chironomidae: Chironominae) of the world.Mem. Am. Entomol. Soc. 36: 1–214.

Grimm, E. C., 1987. CONISS: a FORTRAN 77 program for strati-graphically constrained cluster analysis by the method of incre-mental sum of squares. Comp. Geosci. 13: 13–35.

Grimm, E. C., 1993. TILIAGRAPH 1.25 (computer software). Illi-nois State Museum, Research and Collections Center, Spring-field, II.

Grimm, E. C., 1993. TILIA 2.0 (computer software). Illinois StateMuseum, Research and Collections Center, Springfield, II.

Hallett, D. J., L. V. Hills & J. J. Clague, 1997. New accelerator massspectrometry radiocarbon ages for the Mazama tephra layer fromKootenay National Park, British Columbia, Canada. Can. J. EarthSci. 34: 1202–1209.

Hebda, R. J., 1995. British Columbia vegetation and climate historywith focus on 6 ka BP. Geogr. phys. Quat. 49: 55–79.

Hoffman, R. L., W. J. Liss, G. L. Larson, E. K. Deimling &G. A. Lomnicky, 1996. Distribution of nearshore macroinverte-brates in lakes of the Northern Cascade Mountains, Washington,USA. Arch. Hydrobiol. 136: 363–389.

Hofmann, W., 1978. Analysis of animal microfossils from the GroßerSegeberger See (F.R.G.). Arch. Hydrobiol. 82: 316–346.

Hofmann, W., 1983. Stratigraphy of Cladocera and Chironomidaein a core from a shallow North German lake. Hydrobiologia 103:235–239.

Hofmann, W., 1986. Chironomid analysis. In B. E. Berglund (ed.),Handbook of Holocene Palaeoecology and Palaeohydrology.John Wiley & Sons Ltd., Chichester: 715–727.

Hofmann, W., 1988. The significance of chironomid analysis (Insec-ta: Diptera) for paleolimnological research. Palaeogeogr. Palaeo-climatol. Palaeoecol. 62: 501–509.

Holland, S. S., 1976. Landforms of British Columbia: A Phys-iographic Outline. British Columbia Department of Mines andPetroleum Resources, Bulletin 48, Victoria, 138 pp.

Kansanen, P. H., 1985. Assessment of pollution history from recentsediments in Lake Vanajavesi, southern Finland. II. Changes inthe Chironomidae, Chaoboridae and Ceratopogonidae (Diptera)fauna. Ann. Zool. Fenn. 22: 57–90.

Kansanen, P. H. & J. Aho, 1981. Changes in the macrozoobenthosassociations of polluted Lake Vanajavesi, Southern Finland, overa period of 50 years. Ann. Zool. Fenn. 18: 73–101.

Kansanen, P. H., J. Aho & L. Paasivirta, 1984. Testing the benth-ic lake type concept based on chironomid associations in someFinnish lakes using multivariate statistical methods. Ann. Zool.Fenn. 21: 55–76.

Levesque, A. J., F. E. Mayle, I. R. Walker & L. C. Cwynar, 1993. Apreviously unrecognized late-glacial cold event in eastern NorthAmerica. Nature 361: 623–626.

Levesque, A. J., L. C. Cwynar & I. R. Walker, 1996. Richness,diversity and succession of late-glacial chironomid assemblagesin New Brunswick, Canada. J. Paleolimnol. 16: 257–274.

Lloyd, D., K. Angrove, G. Hope & C. Thompson, 1990. A Guideto Site Identification for the Kamloops Forest Region. BritishColumbia Ministry of Forests, Victoria, 399 pp.

Lotter, A. F., H. J. B. Birks, W. Hofmann & A. Marchetto, in press.Modern cladocera, chironomid, diatom, and chrysophyte cystassemblages as quantitative indicators for the reconstruction ofpast environmental conditions in the Alps. I. Climate. J. Pale-olimnol.

Luckman, B. H., G. Holdsworth & G. D. Osborn, 1993. Neoglacialglacier fluctuations in the Canadian Rockies. Quat. Res. 39: 144–153.

Luckman, B. H. & M. S. Kearney, 1986. Reconstruction of Holocenechanges in alpine vegetation and climate in the Maligne Range,Jasper National Park, Alberta. Quat. Res. 26: 244–261.

Lyons, C. P., 1965. Trees, Shrubs and Flowers to Know in BritishColumbia. J. M. Dent & Sons (Canada) Ltd, Vancouver, 194 pp.

Mathewes, R. W., 1973. A palynological study of postglacial veg-etation change in the University Research Forest, southwesternBritish Columbia. Can. J. Bot. 51: 2085–2103.

Mathewes, R. W., 1985. Paleobotanical evidence for climatic changein southern British Columbia during late-glacial and Holocenetime. In C. R. Harington (ed.), Climate Change in Canada 5: Crit-ical Periods in the Quaternary Climatic History of NorthwesternNorth America. Syllogeus 55: 397–422.

Mathewes, R. W. & L. E. Heusser, 1981. A 12 000-year palynologicalrecord of temperature and precipitation trends in southwesternBritish Columbia. Can. J. Bot. 59: 707–710.

Mathewes, R. W. & M. King, 1989. Holocene vegetation, climate,and lake level changes in the Interior Douglas–Fir Zone, BritishColumbia. Can. J. Earth Sci. 26: 1811–1825.

Miller, R., 1987. Appendix: ash as an insecticide. In B. J. Kemp,Amarna ReportsIV, 14–16. Egypt Exploration Society, London.

Mitchell, W. R. & R. E. Green, 1981. Identification and Interpretationof Ecosystems of the Western Kamloops Forest Region: VolumeII: Dry and Subcontinental Climatic Regions. British ColumbiaMinistry of Forests, Victoria, 223 pp.

Olander, H., A. Korhola & T. Blom, 1997. Surface sediment Chi-ronomidae (Insecta: Diptera) distributions along an ecotonal tran-sect in subarctic Fennoscandia: developing a tool for palaeotem-perature reconstructions. J. Paleolimnol. 18: 45–59.

293

Oliver, D. R., 1968. Adaptations of arctic Chironomidae. Ann. Zool.Fenn. 5: 111–118.

Oliver, D. R., 1971. Life History of the Chironomidae. Ann. Rev.Entomol. 16: 211–230.

Oliver, D. R. & M. E. Roussel, 1983. The Insects and Arachnidsof Canada. Part 11: The Genera of Larval Midges of Canada,Diptera: Chironomidae. Agriculture Canada Publication 1746:263 pp.

Pellatt, M. G., 1996. Postglacial changes in vegetation and climatenear treeline in British Columbia. Ph.D. thesis, Simon FraserUniversity, Burnaby, British Columbia, 177 pp.

Pojar, J., D. Meidinger & K. Klinka, 1991. Concepts. In D. Mei-dinger & J. Pojar (eds), Ecosystems of British Columbia.Research Branch, British Columbia Ministry of Forests, Victoria:9–37.

Porter, S. C. & G. H. Denton, 1967. Chronology of Neoglaciation inthe North American Cordillera. Am. J. Sci. 265: 177–210.

Ryder, J. M. & B. Thomson, 1986. Neoglaciation in the southernCoast Mountains of British Columbia: chronology prior to thelate Neoglacial maximum. Can. J. Earth Sci. 23: 273–287.

Sæther, O. A., 1979. Chironomid communities as water qualityindicators. Holarctic Ecol. 2: 65–74.

Smol, J. P., 1981. Problems associated with the use of ‘speciesdiversity’ in paleolimnological studies. Quat. Res. 15: 209–212.

Thorp, J. H. & R. K. Chesser, 1983. Seasonal responses of lenticmidge assemblages to environmental gradients. Holarctic Ecol.6: 123–132.

Tokeshi, M., 1995. Species interactions and community structure. InP. D. Armitage, P. S. Cranston & L. C. V. Pinder (eds), The Chi-ronomidae: Biology and Ecology of Non-biting Midges, Chap-man & Hall, London: 297–335.

Walker, I. R., 1987. Chironomidae (Diptera) in paleoecology. Quat.Sci. Rev. 6: 29–40.

Walker, I. R., 1988. Late-Quaternary Palaeoecology of Chironomi-dae (Diptera: Insecta) from Lake Sediments in British Columbia.Ph. D. Thesis, Simon Fraser University, Canada, 204 pp.

Walker, I. R., 1990. Modern assemblages of arctic and alpine Chi-ronomidae as analogues for late-glacial communities. Hydrobi-ologia 214: 223–227.

Walker, I. R. & G. M. MacDonald, 1995. Distributions of Chirono-midae (Insecta: Diptera) and other freshwater midges with respectto treeline, Northwest Territories, Canada. Arctic Alp. Res. 27:258–263.

Walker, I. R. & R. W. Mathewes, 1989. Chironomidae (Diptera)remains in surficial lake sediments from the Canadian Cordillera:analysis of the fauna across an altitudinal gradient. J. Paleolimnol.2: 61–80.

Walker, I. R. & C. G. Paterson, 1983. Post-glacial chironomid suc-cession in two small, humic lakes in the New Brunswick-NovaScotia (Canada) border area. Freshwat. Invertebr. Biol. 2: 61–73.

Walker, I. R., J. P. Smol, D. R. Engstrom & H. J. B. Birks, 1991a.An assessment of Chironomidae as quantitative indicators of pastclimatic change. Can. J. Fish. aquat. Sci. 48: 975–987.

Walker, I. R., R. J. Mott & J. P. Smol, 1991b. Allerod – YoungerDryas lake temperatures from midge fossils in Atlantic Canada.Science 253: 1010–1012.

Walker, I. R., A. J. Levesque, L. C. Cwynar & A. F. Lotter, 1997.An expanded surface-water palaeotemperature inference modelfor use with fossil midges from eastern Canada. J. Paleolimnol.18: 165–178.

Wiederholm, T. (ed.), 1983. Chironomidae of the Holarctic Region,Keys and Diagnoses: Part 1 – Larvae. Entomol. scand., Suppl.19: 1–457.