Mosquito larvae (Diptera: Culicidae) in snow-melt pools in a Swedish Lapland area

Transcription

1 June, 2004 Journal of Vector Ecology 109 Mosquito larvae (Diptera: Culicidae) in snow-melt pools in a Swedish Lapland area Christine Dahl 1,3, Lewis T. Nielsen 2, and Erik Petersson 3,4 1 Department of Systematic Zoology, Uppsala University, Uppsala, Sweden 2 Department of Entomology, University of Utah, Salt Lake City, UT, U.S.A. 3 Evolutionary Biology Center, Uppsala University, Uppsala, Sweden 4 Department of Animal Ecology, Uppsala University, Uppsala, Sweden Received 22 July 2003; Accepted 21 December 2003 ABSTRACT: Culicid larvae were collected in snow-melt pools during the first half of June from and in the Abisko Valley, Torne Lapmark, northern Sweden. Twelve species were collected from 102 pools and 81 localities with 7,914 specimens (elevations from 300 to 650 m a.s.l.). Fourteen species are now known from the area. Ten types of larval habitats were classified and ten species were statistically analyzed on pooled values over all years for abundance, coexistence, and diversity. The greatest abundance of individuals had pools with Carex spp. on the bottom. Most widespread and abundant were Oc. hexodontus (Dyar), Oc. communis (DeGeer), Oc. pullatus (Coquillett), and Oc. punctor (Kirby). The two circumpolar, arctic species, Oc. nigripes (Zetterstedt) and Oc. impiger (Walker), were mainly confined to elevations above 400 m a.s.l. In such marginal habitats as these pools, exposed to variable abitotic conditions, the phenology of species was rather constant. Species diversity varied somewhat between habitats. Oc. hexodontus, Oc. communis, and Oc. pullatus most often occupied pools as single species. Oc. communis and Oc. hexodontus had the highest values for coexistence and also occurred with Oc. punctor. The habitats that were richest in species were those without vegetation or with detritus on the bottom and surrounded by Empetrum nigrum and Betula nana. These habitats contained between eight and ten species. Species abundance and coexistence over the years showed no stable species patterns for the habitat types in the area. Journal of Vector Ecology 29 (1): Keyword Index: Mosquito larvae, distribution, ecology, Abisko. INTRODUCTION The huge numbers and persistent biting habits of female mosquitoes in the Abisko area (Lapponia, Sweden) are experienced by many visitors. This phenomenon had been commented upon even before Linnaeus visited Lapponia in He succinctly mentioned in the chapter about his Jokkmokk journey that mosquitoes were troublesome (Linnaeus 1975). Thienemann (1938a,b) cites from Linnaeus Flora Lapponica Culex vulgaris (a nomen dubium, Knight & Stone, 1977, probably a Ochlerotatus sp.) as an obnoxious biter. This well-known problem of mosquito mass occurrence in northern Sweden was first examined by Thienemann (1938b) during his studies of northern Chironomidae. Records of biting females from the Abisko area have been published by Edwards (1931), Thienemann, who also collected larvae (1938a,b), Dahl (1974a), and by Nielsen (1979). Nectar feeding by adult mosquitoes was studied near Abisko by Andersson (1991) and a single year s trap collection was analyzed by Schäfer ( unpublished data). Natvig (1948) recorded adults and larvae in adjacent areas and Dahl (1974b) compiled records from Lule Lapmark. No long-term overview of species distribution in larval habitats in the valley and on adjacent mountain slopes exists prior to the study reported here. Thienemannn (1938b) published a two-year study of two pools near the research station. In an area where weather can vary drastically in one year as well as between years, it is particularly interesting to have series of collections from the same habitats over several years. We initiated such an investigation to establish a basis for further studies of female mosquito movements in the Abisko Valley, which are still being carried out.

2 110 Journal of Vector Ecology June, 2004 MATERIALS AND METHODS Larvae were collected in the vicinity of the Abisko Scientific Research Station of the Royal Swedish Academy of Sciences (ANSR, 68 o 19 N 18 o 45 E) in the Abisko Valley, Torne Lapmark, in northwestern Sweden. This area includes a portion of the Abisko National Park and a larger area near Lake Torne Träsk, m a.s.l. to slopes between 600-1,050 m a.s.l. of the surrounding mountains with tops to S and SE of 1,400-1,700 m, to NW 1,000-1,500 m. The valley opens NE and E toward Lake Torne Träsk. The Abiskojokk and Nissanjokk rivers are the most prominent rivers dividing the area. Collections were made during mid-june ( ) in 1975, 1976, 1977, 1979, 1995, and A total of 102 pools at 81 localities were sampled (Figure 1). Approximately one-fourth (28/102) of the sites were sampled more than once, either in the same year or in different years, resulting in 253 site visits between Ten dips per pool were taken with a 350 ml dipper at all 102 pools that were sampled (Figure 1). During the last two years, mainly the higher elevations and the area east of the Nissunjåkka and Kallolako Rivers were sampled because most of the pools in the valley had already dried out. Larger pools were sampled along transects. An interval of a few minutes between dips allowed disturbed and diving mosquitoes to return to the water surface. Pool size varied greatly between habitats and between years, depending on the extent of snow, precipitation, and rapidly changing weather conditions. Water depth and approximate pool sizes were recorded for most samples in the first year, but these data were not included in analyses because of heavy fluctuations between years. General weather conditions for the relevant periods were obtained from the Abisko Scientific Station s weather reports (Table 1). Elevation and configuration of the vegetation in and around pools were duly noted and ten habitat types were classified based on the presence, absence, and type of vegetation: Class 1) presence of dominating vegetation in the pools A-D. A: Carex spp.; B: grasses; C: Sphagnum spp.; D: lake shores with Carex spp. on inundated edges. Class 2) absence of vegetation in the pools E-J. Two subclasses were identified. Stony pools (E, F); E: stony pools in Sphagnum areas, F: stony pools in Empetrum nigrum L. and Betula nana L. areas both with little detritus between stones at the bottom. Pools with rich detritus and leaves at the bottom (G - J); G: detritus rich bottom in Sphagnum areas, H: detritus rich bottom in Empetrum, Betula nana areas, I: detritus rich bottom in Salix spp and/or Betula pubescens Ehrh. (Tutin et al. 2000), J: detritus rich bottom surrounded by Trollius europaeus L. and high grass. Although all larval instars and pupae were collected, only about 70% of the material was identified. This consisted of instars III and IV, for which identification was certain. However, instar I and II larvae from pools in which only one species was present were included in the analyses. For pools with several species present, the early instar larvae were used for an approximate overall assessment of phenology. Otherwise, only identified material to species was included in summaries and statistical analyses. Occasional netting and rearing of larvae to adults (especially during 1978) and an unpublished human landing rate study conducted by the first author in 1976 supplemented species determinations. For statistical analysis, Shannon-Wiener indices (Margalef 1958) were used to assess species diversity (Lloyd and Ghelardi 1964). Yule s V indices were used to evaluate species association (Pielou 1977). A list of localities is available at ANSR. Larval material and reared adults are deposited at the Zoological Museum, Lund University, Sweden. Ochlerotatus is used as the genus name in accordance with Reinert (2001). RESULTS Distribution of species Based on our data (Figures 1 and 2), and previous records (Table 2), 14 species of Culicidae are now known from the Abisko area either as larvae or adults or both (Thienemann 1938a,b, Natvig 1948, Dahl 1977, Nielsen 1985, Schäfer (unpublished data), Schäfer and Lundström 2001). The genus Ochlerotatus was dominant in this area with eleven species including the new record of Oc. intrudens, found as larvae at the slope of the Abisko Valley and in the area of Stordalen (loc. 198) as adults. Larvae of Oc. pionips (Dyar) and Oc. hexodontus (Dyar) also had not been recorded previously. Of the three Culiseta species, Cs. morsitans Theob. larvae were new and those of Cs. bergrothi Edw. have not yet been found. From Stordalen east of Abisko (several localities and different habitats), we report Oc. punctor (Kirby), Oc. punctodes (Dyar), Oc. communis, and Oc. hexodontus larvae. Over the Abisko Valley the most widely distributed species were Oc. communis, Oc. hexodontus, Oc. pullatus (Coquillett), Oc. pionips (Dyar), and Oc. punctor (Figure 1). Distribution relative to elevation also varied between species. Larvae of all common species occurred below 400 m a.s.l., but rarely those of Oc. punctodes, Oc. intrudens, and Cs. alaskaensis (Ludlow). Both Oc. nigripes (Zetterstedt) and Oc. impiger (Walker) are subarctic to high arctic, circumpolar Ochlerotatus species (Danks, 1981). In the Abisko area substantial numbers of these species were

3 June, 2004 Journal of Vector Ecology 111 Figure 1. Distribution of our material (12 species) as larvae in the Abisko Valley.

4 112 Journal of Vector Ecology June, 2004 Table 1. Air temperature dates, amount of snow cover per month, and first ice free day of Torne Träsk lake, compiled from ANSR weather station records. Year C Max Min Snow cover Icefree Torneträsk 1975 April > -9.0 all month May days June none, some topsnow 10 June July none 1976 April all month May days June day 11 June July none 1977 April < -10 all month May days June none 17 June July none 1979 April > days May days June none 13 June July none 1995 April > all month May days June none 10 June July none 1996 April all month May days June none 14 June July none

5 June, 2004 Journal of Vector Ecology 113 Figure 2. Phenology of species during our two sampling periods, per year and habitat. # not sampled; * not present in samples. found only in larval habitats around and above m a.s.l., i.e., above the tree line. Phenology Weather varied considerably throughout the annual sampling periods. During the main collecting period in June, the average monthly temperatures were much higher than those of the previous months, despite the less variable snow-melt and ice-free dates (Table 1). The sampling period covered the phenology of all larval instars to pupae. Two subperiods enabled interspecific comparisons of the phenologies of instars III - IV between species, independent of years, habitat class, or height above sea level (Figure 2). This analysis showed that three of the most frequent and abundant species, Oc. communis, Oc. hexodontus, and Oc. punctor, have a rather consistent phenology. The two most abundant species, Oc. communis and Oc. hexodontus, present in the III and IV instars during both subperiods (Figure 2, Table 3), had a similar phenology despite variable air temperatures between years in late May or early June. A comparison between years, species, instars, and the pupal stage indicated that the five most abundant species in this period also were the earliest ones to hatch. From 1975 to 1977, these species had a slight tendency for earlier eclosion, whereas Oc. impiger, Oc. excrucians, and Oc. nigripes showed more variation between years. The years 1979, 1995, and 1996 were slightly warmer (Table 1). The localities of Oc. pionips ( ), and Oc. impiger (1995) had dried out. In those years some pools already were dry in mid-june, while others contained mainly unidentified pupae. Thus, weather conditions affected to some extent the amount of individuals sampled and identified for each year. Habitats, species abundance, and frequency Specimen abundance and frequency in the different habitat classes is summarized in Figure 3. The most

6 114 Journal of Vector Ecology June, 2004 Figure 3. Pooled numbers of ten species, III- IV instars, according to ocurrence in habitats. For habitat designations and explanations see Table 4.

7 June, 2004 Journal of Vector Ecology 115 Table 2. Records of species and authors from the Abisko area. See literature references for author abbreviations. Ed.1931 Thi.1938 Nat.1948 Da.1974a An.1991 Sch.(unpub.) Our study Oc. communis (DeG.) X X X X X X Oc. diantaeus (HDK) X X Oc. excrucians (Walker) X X X X X X X Oc. hexodontus (Dyar) X X X X Oc. impiger (Walker) X X X X X X Oc. intrudens (Dyar) X X Oc. nigripes (Zetterstedt) X X X X X Oc. pionips (Dyar) X X X Oc. pullatus (Coquillett) X X X X X X Oc. punctodes (Dyar) X X Oc. punctor (Kirby) X X X X X X X Cs. alaskaensis (Ludlow) X X X X Cs. bergrothi (Edwards) X Cs. morsitans (Theobald) X common species occurred with the highest abundance of individuals in habitats with vegetation (A) and also in habitats without vegetation but detritus at the bottom and surrounded by vegetation (G, H, I). Partly they were found in flat pools with Carex and partly in stony pools or in somewhat deeper but smaller pools with only detritus at the bottom (Table 4). There existed a difference between Oc. pionips with larvae most abundant below 400 m a.s.l. and Oc. pullatus with larvae more abundant at higher levels up to 600 m a.s.l. Oc. nigripes at its lower limit (450 m a.s.l.) preferred deeper pools with detritus rich bottom and overhanging, low vegetation. It was also found some years at a small lake 500 m a.s.l. in a rather shallow pool with Carex spp. At higher elevations the species preferred pools with Carex, or in high mountainous areas they preferred flat pools with Sphagnum and/or Polytrichium spp. The larvae were easily disturbed, dived quickly, and stayed at the bottom for a longer time. Oc. impiger occurred in great numbers in small and large, relatively shallow pools with detritus near and above 400 m a.s.l. (Figure 1), most abundantly near the tree line where the pools are often surrounded by Salix spp. and/or Betula nana with rich detritus bottom. The abundance of species, between nine and eight species, was rather similar in habitats A, G, F, and H, if the number of samples was considered. The other habitats yielded between seven and six species. Habitats B and D had five species and J only two species (Table 4). The highest species frequency in different habitats had Oc. hexodontus, Oc. communis, and Oc. punctor, whereas Oc. pullatus and Oc. pionips had lower values. The rareness of Oc. intrudens, Oc. punctodes, and Cs. morsitans larvae was evident (Table 5). Statistical analyses of coexistence and species diversity An initial analysis between species for coexistence showed a high coincidence between Oc. communis and Oc. hexodontus larvae. Both species often occurred with Oc. pionips, Oc. pullatus, and Oc. punctor. When species were analyzed for coexistence on pooled values and independent of habitat classes, we found that Oc. hexodontus had the highest solitary presence closely followed by Oc. communis and Oc. pullatus. (Table 5). When analyzed statistically by Yule s V test (Pielou 1977), which is based on a paired comparison for association (Table 6), only a few significant values were obtained. Out of 45 unique species pair combinations, only four had significant values: Oc. excrucians - Oc. impiger, Oc. impiger - Oc. punctor, Oc. nigripes - Oc. punctor, and Oc. pionips - Oc. punctodes. However, Oc. punctodes and Oc. nigripes were infrequent in the samples, and therefore these values were not considered further. The combinations of the two first pairs showed only rare associations as significant. Species diversity was analyzed by proportional abundance of species within habitat types. Values were

8 116 Journal of Vector Ecology June, 2004 Table 3. Phenology of eight species in three groups (I-II and III-IV instars) during two periods of sampling per year. Years/spp. Oc. communis Oc. hexodontus Oc. pullatus Oc. pionips Oc. punctor Oc. impiger Oc. exrucians Oc. nigripes Date /Instars I/II(III+IV) I/II (III+IV) I/II (III+IV) I/II (III+IV) I/II (III+IV) I/II (III+IV) I/II (III+IV) I/II (III+IV) /9 (452) 0/5 (116) 0/59 (132) 0/38 (26) 0/8 (81) 0 /1 (9) 0/1 (5) - / - (-) /87 (340) 0/2 (52) 0/46 (40) 0/71 (71) 0/0 (61) 0/0 (8) 0/0 (0) - / - (-) /0 (1) 0/0 (11) 0/1 (3) 0/1 (7) 0/0 (1) 0/2 (6) 0/0 (0) 0/0 (0) /2 (32) 0/6(70) 0/5 (25) 0/0 (1) 0/0 (4) 7/0 (0) 0/0 (4) 0/0 (4) /47 (237) 2/23 (117) 1/9 (22) 0/0 (6) 0/21 (5) 3/11 (12) - / - (-) - / - (-) /4 (142) 0/10 (92) 2/5 (12) 0/2 (19) 0/4 (37) 0/0 (1) - / - (-) - / - (-) /0 (106) 0/2 (101) 0/0 (99) 0/0 (51) 0/0 (59) 0/0 (4) 0/2 (17) 0/0 (1) /3 (400) 0/0 (89) 0/1 (85) 0/0 (135) 0/7 (48) 0/0 (18) 0/0 (170) 0/1 (45) /9 (30) 0/1 (69) 0/34 (33) - / - (-) 0/2 (11) - / - (-) 0/0 (1) 0/0 (1) /1 (15) 0/1 (22) 1/10 (0) - / - (-) 0/5 (7) 0/0 (-) 0/0 - / - 0/0 (2) /5 (27) 0/4 (26) 0/0 (7) - / - (-) 0/2 (50) 0/4 (10) 0/2 (-) - / - (-) /10 (27) 0/1 (59) 5/26 (22) - / - (-) 2/6 (22) 2/20 (90) - / - (-) - / - (-)

9 June, 2004 Journal of Vector Ecology 117 Table 4. Ten habitat classes represented by number of pools, number of visits, number of larvae (III-IV instars) and pupae in the pools, and number of species found in the habitats. Habitat Class n/pools n/visits s/spec. n/spp mean n/sp n/ With vegetation at bottom A Carex spp B grasses C Spagnum sp D lake shores Without vegetation at bottom E stony, Sph F stony, E,Bn G detritus, Sph H detritus, E, Bn I detritus, Salix, B J detritus, Trollius pooled over the years (Table 7). In that analysis, habitats G and A had the highest diversity indices (H*), followed by habitats H, F, and C. The highest maximum diversity values had A and H (H max ). Thus, the Carex and detritus habitats were richest in species. When overall homogeneity of species (Table 7) was analyzed, habitats A, B, C, and H had the most homogenous species configurations. Specific species showed differences in regard to proportional abundance in habitats. Oc. communis had its highest abundance in habitats G, H, F and A; Oc. hexodontus in D, C, B and E; Oc. pullatus in E and F; and Oc. punctor in E, A, and H. The high values of Oc. impiger in B and Oc. pionips in J depend on their solitary occurrence in one pool of these habitats. DISCUSSION Diversity and distribution Comparison of species diversity between regions in Sweden was made by Schäfer and Lundström (2001). We expand the overview here and our results with fourteen species from the Abisko Valley represent the highest species diversity from a restricted northern Fennoscandian locality. From Norway Inner Troms, west of Torne Lapmark, Mehl (1996) recorded six of the species found in Abisko plus Ae. cinereus Meigen. This species and Ae. vexans (Meigen) were not previously found in the Abisko area. Twelve species have been recorded as adults from Allavaare, Lule Lapmark (Schäfer and Lundström 2001). Utrio (1984) recorded eleven species, including Oc. beklemishevi Den. but not Oc. nigripes and Oc. impiger from Inari Lapland in north Finland. He suggested a change in biodiversity towards Kemi Lapland with sixteen species recorded, including Oc. diantaeus H., D. & K., Oc. nigrinus Eckst., Oc. riparius Dyar & Knab, Ae. cinereus Meig., and An. maculipennis s.l. Meig. Our conclusion is that local distribution patterns of species from traps in Fennoscandian Lapland areas depend so much on altitude, ground conditions, and general weather conditions that some variations as to species numbers and composition can occur over fairly small areas. Seven of the species found by us as larvae in Abisko were caught as adults in barns and stables in Västerbotten and Norrbotten Län and Lule Lapmark (Jaenson 1985). Diversity in Swedish mosquitoes is steadily increasing towards the southern regions. From Lule Lapmark seventeen species (Dahl 1974b), from Sässman, Hälsingland, seventeen species (Jaenson et al. 1986), from Uppland twenty-three species (Jaenson 1990), and from Scania thirty-two species (Dahl 1977) have been reported. To get an overview of species diversity in Sweden, Schäfer and Lundström (2001) established fourteen functional groups on 41 species (own samples and literature records). These functional groups were based

10 118 Journal of Vector Ecology June, 2004 Table 5. Pooled values of coexistence for eleven species. Bold numbers indicate the solitary existence of one species in a pool. Oc. Oc. Oc. Oc. Oc. Oc. Oc. Oc. Oc. Oc. Cs. com exc hexo impiger intr nigr pion pull punc punct mors Oc. communis Oc. excrucians Oc. hexodontus Oc. impiger Oc. intrudens Oc.nigripes Oc. pionips Oc. pullatus Oc. puntor Oc. punctodes Cs. morsitans on a mixture of life traits in eggs, larvae (egg hibernation versus larval hibernation, univoltism or multivoltism), and adults (adult hibernation, mixed mammal blood sources versus bird blood, or preference for amphibian blood) and were mixed with single characteristics in egg laying patterns, habitat selection, or morphology. From Abisko, they recorded three functional groups (1, 7, 8). Species from their functional group 7 were the most abundant ones in our samples. Local ecological variations in tundra and high mountainous species were many. Our results of larval distribution in the Abisko area showed only univoltine species but also some of the resilience of female behavior and differences in life traits in northern areas. Females of Oc. impiger and Oc. nigripes prefer to lay their eggs above the birch tree line but enter the lower areas of the valley in search of blood meals and occasional egg laying, and they thus broaden a condition of Schäfer and Lundström s functional group 7. The few Culiseta species found in the northern mountains are not multivoltine, thus violating one of the conditions of group 1. The circumpolar Oc. hexodontus was one of the most abundant and frequent species in our material and exhibits traits of egg-laying and biting within birch forest, mixed forest, and above tree limit localities with still more differences recorded from other high mountainous areas with mixed forests in the Nearctic (Nielsen and Blackmore 1996). Important questions present themselves as to how functional groups of insects with different life stages should be defined. Can single characteristics, ecological or morphological, be mixed with life traits as the basis for definition of functional groups? Which criteria will be most relevant for all geographical regions where the species can be found? In our view definitions of ecological functional groups in insects should adhere much more strictly to one life stage and to special and well-defined categories of ecological traits of general relevance. Our analysis of the distribution of larval populations over several years in one area and many habitats should have been combined with adult trap collections in the same area and the same years to indicate female dispersal in common species, but this was not done. However, for known restricted distributions of rare species, our comparison with the results of human baits and trap collections from 1996 (Schäfer, unpublished data) revealed some general trends. If the restricted larval distribution of Oc. impiger and Oc. excrucians over the years (Figure 1) is compared with trap collections of 1996 (Schäfer, unpublished), Oc. impiger females seem to use mass dispersal for a blood meal into the lower parts of the valley, whereas Oc. excrucians females exhibit a reverse search pattern. Ochlerotatus nigripes and Oc. impiger, which occur

11 June, 2004 Journal of Vector Ecology 119 Table 6. Significance values of coexistence analysis by pairwise association (Yule s index). Bold numbers indicate significant values. Oc. Oc. Oc. Oc. Oc. Oc. Oc. Oc. Oc. excrucians hexodontus impiger intrudens nigripes pionips pullatus punctor punctodes Oc. communis (0.127) (0.697) (0.022) (0.398) (0.198) (0.573) (0.955) (0.298) (0.085) Oc. excrucians (0.437) (0.463) (0.747) (0.938) (0.963) (0.611) (0.932) (0.268) Oc. hexodontus (0.940) (0.544) (0.798) (0.284) (0.210) (0.147) (0.938) Oc. impiger (0.130) (0.492) (0.215) (0.307) (0.022) (0.389) Oc. intrudens (0.693) (0.517) (0.333) (0.398) (0.811) Oc. nigripes (0.960) (0.587) (0.031) (0.481) Oc. pionips (0.071) (0.111) (0.039) Oc. pullatus (0.143) (0.267) Oc. punctor (0.085)

12 120 Journal of Vector Ecology June, 2004 circumpolar in subarctic and arctic areas, have not been found in European high mountainous areas, and only one record exists of Oc. hexodontus from the Alps (Dahl and Nielsen 1994). A record of Oc. nigripes near the polish baltic coast (Skierska 1962) pictures the larva of another species and its identification is thus doubtful. Distribution in Europe of over 100 species has been dealt with by Snow and Ramsdale (1999), Schaffner et al. (2001), and Becker et al. (2003). Nearctic distribution of species recorded from Abisko is well-documented (Nielsen and Blackmore 1996). From the Palearctic, the Siberian Tundra, and other northern regions, distributional, ecological, and life trait data of all stages of these species are still awaiting compilation. Phenology Pooled data of the three common species Oc. communis, Oc. hexodontus, and Oc. punctor showed little phenological variation between years (Figure 3). Pooling will override individual differences in larval phenology at various heights or between various years. For predictions of mass occurrence of a mixture of these species our approach was sufficient. For predictions of which species will be the most common biters throughout the day during a certain year, more detailed investigations are needed. The mosaic of species during several years in a known locality is an example. Thienemannn (1938b) found Oc. communis, Oc. punctor, and a few Oc. impiger in his Mückentümpel (our locality 185). He also had an ecological explanation for the mass occurrence of mosquitoes in these northern localities. He considered the layer of frost at 1-2 m below the ground during 1938 as an effective barrier to sinking of snow-melt water. During the years 1975 to 1995 we found Oc. communis, Oc. pionips, Oc. pullatus, and Oc. punctor together or alone but never Oc. hexodontus or Oc. impiger in locality 185. In 1996 the Mückentümpel (locality 185) held 10 cm water depth and Oc. communis larvae only. In the years 1979 and 1995 the pool had dried out completely during our sampling period. Now, after a road improvement, the locality has been partly destroyed. When several years and two different ground configurations (morän and peat) were considered in this locality, the picture became more complicated (Table 1 and ground frost records, courtesy of ANSR, Abisko Scientific Research Station). Peat ground in our sampling years and time never held frost at 1.5 to 2 m below ground and thus a lot of our snow-melt pools were situated on frost-free ground. When in June 1976 and 1977 morän ground frost occurred at 1.5 to 2 m below ground during May, locality 185 dried out very early. Our material had too much variance in abiotic factors to fully explain species differences between localities over different Table 7. Species diversity index according to Shannon-Wiener function. Proportional abundance of species in ten habitats. H* = highest diversity index; H max = highest maximum diversity; Equit. = highest homogenous species configuration. Habitat Oc. Oc. Oc. Oc. Oc. Oc. Oc. Oc. Oc. Oc. H* H max Equit. communis exucians hexodontus impiger intrudens nigripes punctodes pionips pullatus punctor A B C D E F G H I J

13 June, 2004 Journal of Vector Ecology 121 years. Little snow during some winters (Table 1), pool size and depth (recorded in some years by us), and warm weather in May together with low minimum temperatures during that period will influence the amount of water in snow-melt pools during larval development. Under certain weather conditions pools might dry out in early June. Nilsson and Svensson (1994), in a more southern locality, found that higher temperatures could cause mortality of small Oc. communis instars. Insolation is an important factor for rapid larval development in the Abisko area (Thieneman 1938a, b) and could influence species abundance or even cause shifts in species presence over the years. Differences of egg and larval survival in flat pools in marginal, northern habitats, or unknown mechanisms of egg-laying behavior in females may form transient species combinations according to weather conditions. These problems remain to be studied. Habitat classes, species frequency, and coexistence patterns Our classification of habitats is much more detailed than the international category, above ground waters, in Laird (1988) or the string bog habitat investigated by Maire (1982). Habitats in the Abisko area are associated with lake edges, swamps, marshes, permanent ponds, and temporal, snow-melt pools. The latter comprised the highest number of localities we sampled. Our more detailed categorization (Table 4) gave a better picture of the habitat mosaic in the area. The habitats differed slightly in species diversity when pooled and analyzed for diversity by the Shannon-Wiener index (Table 7). Most habitats had a diversity index of more than two over the years, whereas habitats D, I, and J had lower indices. Lake shores (D) were poor in both species and specimens. Stony pools without vegetation on the bottom and with surrounding Sphagnum (G) had an overall presence of Oc. communis, and the occurrence of Oc. pionips and Oc. punctor brought this habitat into the highest diversity group. This might be a result of pooling and sampling at different elevations. The detritus-rich pools with Salix and Betula (I) of Oc. impiger at higher elevations, their main habitat, had a low species index. The low values for Oc. pionips in habitat (J) were caused by the one time, two-year sampling. This species was only found in greater numbers in habitats near Lake Torne Träsk in the first four years. During 1995 and 1996 its main localities there were dried out. Thus, local conditions of these temporary pools can skew diversity results on pooled values to a considerable extent. Diversity maximum values in our analysis also reflected the higher number of samples. Equity values were similar for habitats A, B, C, and G and differed in the other habitats. A large variation in number of individuals was encountered in and between different habitats (Tables 3 and 4). Abundance and coexistence was analyzed by two methods on pooled values. The results of Yule s V association index (Table 6) showed that out of 45 unique species pair combinations only four were significant (Oc. excrucians - Oc. impiger, Oc. impiger - Oc. punctor, Oc. nigripes - Oc. punctor, and Oc. pionips- Oc. punctodes). However, Oc. punctodes and Oc. nigripes were infrequent in the samples, and therefore these significant values should be regarded as anecdotal. The other two pairs show only rare associations as significant. The number individuals/pool differed between habitats (χ 2 =400.9, d.f.=9, P<0.001), as well as between type of habitats (with or without vegetation at the bottom (χ 2 =70.7, d.f.=1, P<0.001)). However, the number of species/pool did not differ between habitats (χ 2 =6.53, d.f.=9, P=0.68) or between type of habitat (χ 2 =0.56, d.f=1, P=0.457). The numbers of species in the pools were correlated with the numbers of individuals (r s 0.761, n=10, P=0.0106), indicating that more species are found in pools that can supply many individuals. The patchy species distribution and the lack of stable species associations in habitats most likely reflect the unpredictable situations in these temporary pools. Some years the pools contain much water, being refilled with both snow-melt water and rain throughout the growth period of the larvae. There are probably also differences of water influx between the pools. If most of the water is derived from melted snow, some deeper and larger pools will be supplied with more water than smaller and flatter ones. In a warm spring, before or during larval growth, flat or small pools may dry out relatively early. Such unpredictable variations in spring within and between pools may play an important role in diminishing the probability of stable associations of species over many years. Feeding modes of Ochlerotatus species, such as filter feeding and scraping (Dahl 1988), may in general favor their occurrence in deeper and stony pools. Surface feeders like Anopheles larvae might find less favorable conditions in oligotrophic pools without surface films (Timmermann and Briegel 1993). The patchy and rare occurrence of a few predators, such as Mochlonyx sp. larvae, might be a consequence of the instable conditions in such northern habitats. In regions where habitat size (Sunahara et al. 2002), insolation, and predator diversity (Nilsson and Svensson 1994) are stable, culicid species associations will be easier to find. In conclusion, qualitative sampling of larvae over years in marginal habitats in an area where weather conditions change quickly and heavily will give broad results for species distribution, diversity patterns, larval abundance, and species frequency in specified habitats.

14 122 Journal of Vector Ecology June, 2004 By sampling precise and corresponding time windows, our results could be used to trace shifts in habitat presence over time and showed that our habitat classification was adequate. Our results on larval distribution suggest that a locally detailed habitat categorization enables understanding of the problems of apparent variability in survival of eggs or small instar larvae under shifting abiotic conditions. It also contributes to the understanding of how single species move over an area to find blood sources and how to classify northern larval habitats. Our scheme should be applied to similar biotopes north of the Arctic Circle or in other northern palearctic regions. Such comparisons of larval studies in restricted areas would contribute to the understanding of mass occurrence of females and the ecological demands and behavioral traits of northern holarctic culicid species. Acknowledgments The staff of the Abisko Scientific Research Station of the Royal Swedish Academy of Sciences (ANSR) has been very supportive throughout our investigation. We especially thank Nils Åke Andersson for his help on practical matters and retrieving meteorological data. Astrid Ulfstrand, Uppsala, kindly drew the distribution map of the Abisko area. Our colleague Jan Bengtsson, SLU, provided valuable comments on one of the earlier drafts. We also thank Mark Blackmore, Valdosta, for valuable comments and corrections of our English. Financial support was provided by The National Geographic Society of America to L. T. Nielsen and by the Nils von Hofsten Foundation, Uppsala University and the Magnus Bergvall Foundation, SEB Sweden to C. Dahl. REFERENCES CITED Andersson, H Nectar feeding behaviour and the significance of sugar meals in mosquitoes (Diptera: Culicidae). Acta Univ. Upsaliensis* 358: *Comprehensive Summaries of Uppsala Dissertations from the Faculty of Sciences. Becker, N., D. Petric., M. Zgomba, C. Boase, C. Dahl, J. Lane, and A. Kaiser Mosquitoes and their control. Kluwer Academic/Plenum 500 pp. Dahl, C. 1974a. Circumpolar Aedes (Ochlerotatus) species in North Fennoscandia. Mosq. Syst. 6: Dahl, C. 1974b. Stickmyggor (Diptera, Nemat.: Culicidae) i Messaure området. Norrbottens Läns Naturvårdsförbund 30 (Norrbottens Natur Småskrift) 1: Dahl, C Taxonomy and geographic distribution of Swedish Culicidae (Diptera, Nematocera). Entomol. Scand. 8: Dahl, C Control potential in feeding mechanisms of mosquito larvae. Bull. Soc. Vector Ecol. 13: Dahl, C. and L.T. Nielsen Hochalpine Stechmücken aus Österreich (Diptera: Culicidae). Entomol. General. 18: Dahl, C Diptera Culicidae, Mosquitoes. In: (A. Nilsson (ed.) Aquatic Insects of North Europe. A Taxonomic Handbook Vol 2. pp Apollo Books, Stenstrup, Denmark. Danks, H.V Arctic arthropods: A review of systematics and ecology with particular reference to the North American fauna. Tyrell Press, Ottawa. 608 pp. Edwards, F.W Myggor. In: Y. Sjögren (ed.) Insektfaunan inom Abisko Nationalpark. III. pp K. Svenska Vetenskapsakademiens skrifter i Naturskyddsärenden 18. Jaenson, T.G.T Stickmyggor - Aedes och Culiseta - i norrländska ladugårdar och stallar. Ent. Tidskr. 106: Jaenson, T.G.T., B. Niklasson, and B. Henriksson Seasonal activity of mosquitoes in an Ockelbo disease endemic area in central Sweden. J. Am. Mosq. Contr. Assoc. 2: Jaenson, T.G.T Vector roles of fennoscandian mosquitoes attracted to mammals, birds and frogs. Med. Vet. Entomol. 4: Laird, M The natural history of larval mosquito habitats. Academic Press. San Diego, CA. 555 pp. Linnaeus, C Lapplandsresa år M. Platen and C-O. Sydow, eds. (in Swedish, annotated). Lloyd, M. and R.J. Ghelardi A table for calculating the equitability component of species diversity. J. Anim. Ecol. 33: Maire, A Selectivity by six snow-melt mosquito species for larval habitats in Quebec subarctic string bogs. Mosq. News. 42: Margalef, R Information theory in ecology. Gen. Syst. 3: Mehl, R Culicidae Stickmygg. In: K. Aagard and D. Dolmen (eds.) Limnofauna Norvegica: Tapir forlag Tronheim. Natvig, L.R Contributions to the knowledge of the Danish and Fennoscandian mosquitoes. Norsk Entomol. Tidskr. Suppl. I. Oslo. 567 pp. Nielsen, L.T A comparative taxonomic and distributional study of holarctic Aedes mosquitoes. Natl. Geogr. Soc. Res. Rpt. 20: Nielsen, L.T. and M.S. Blackmore The mosquitoes

15 June, 2004 Journal of Vector Ecology 123 of Yellowstone National Park (Diptera: Culicidae). J. Am. Mosq. Contr. Assoc. 12: Nilsson, A.N. and B.W. Svensson Dytiscid predators and culicid prey in two boreal snowmelt pools differing in temperature and duration. Ann. Zool. Fennici 31: Pielou, E.C Mathematical ecology. Wiley- Interscience. New York 385 pp. Reinert, J.F Revised list of the abbreviations for the genera and subgenera of Culicidae (Diptera) and notes on generic and subgeneric changes. J. Am. Mosq. Contr. Assoc. 17: Schaffner, F., G. Angel, B. Geoffroy, J.-P. Hervy, A. Rhaiem, and J. Brunhes Les moustiques d Europe. IRD edition (CD ROM). Paris, France. Schäfer, M. and J.O. Lundström Comparison of mosquito (Diptera: Culicidae) fauna characteristics of forested wetlands in Sweden. Ann. Entomol. Soc. Am. 94: Skierska, B Discovery of the subarctic mosquito species Aedes (Ochlerotatus nigripes Zetterstedt (Culicidae) in Poland. Bull. Inst. Marine Medic. in Gdansk1/2:XIII: Snow, K. and C. Ramsdale Distribution chart for European mosquitoes. Eur. Mosq. Bull. 3: Sunahara, T., K. Ishizaka, and M. Mogi Habitat size: A factor determining the opportunity for encounters between mosquito larvae and aquatic predators. J. Vector Ecol. 27: Thienemann, A. 1938a. Frostboden und Sonnestrahlung als limnologische Faktoren. Arch. Hydrobiol. 34: Thienemann, A. 1938b. Die Ursachen der Stechmücken- Plage im hohen Norden. Natur und Volk. Senckenbergische Naturforschende Gesellschaft, Frankfurt a.m. 68: Timmermann S.E. and H. Briegel Water depth and larval density effect development and accumulation of reserves in laboratory populations of mosquitoes. Bull. Soc.Vector Ecol. 18: Tutin, T.G, N.A. Burges, A.O. Chater, J.R. Edmondson, V.H.Heywood, D.M. Moore, D.H. Valentine, and S.M. Walters Flora Nordica.vol.1. (Ed. B. Jonsell) p.201. The Bergianus Foundation. The Royal Swedish Academy of Sciences, Stockholm, Sweden. Utrio, P Culicidae (Diptera) of Inari Lapland. Invertebrates of Inari Lapland Kevo Notes 7: