key: cord-0009910-lc8712xm authors: HODKINSON, I. D.; JENSEN, T. S.; MacLEAN, S. F. title: The distribution, abundance and host plant relationships of Salix‐ feeding psyllids (Homoptera: Psylloidea) in arctic Alaska date: 2008-03-14 journal: Ecol Entomol DOI: 10.1111/j.1365-2311.1979.tb00568.x sha: c00007b831e0fc8d9b1105090c1efaf6da142bc6 doc_id: 9910 cord_uid: lc8712xm Abstract. 1. Five species of psyllid occurred on seven species of Salix at Meade River, Alaska. Studies were made on the two common species Psylla pclmeni Löw and P.phlebophyllae Hodkinson. The former feeds on the phanerophy tes Salix pulchra, S.lanata, S.alaxensis and S.glauca, the latter on the chamaephytes S.phlebophylla and S.reticulata. 2. Both P.palmeni and P.phlebophyllae had a 1‐year life cycle and nymphal development took place on the female Salix catkin. The life cycle was generally closely synchronized with the period of catkin development. However, only a few eggs were laid on S.glauca 3. Seasonal perturbation of the host plant by flooding, ice movement and blown sand prevented psyllids breeding in certain areas colonized by the host plant. 4. In P.palmenidensities and ‘feeding pressure’, measured as biomass of psyllids per gram of catkin, on the different host plants followed the sequence S.pulchra>S.lanata> S.alaxensis > S.glauca. In P.phlebophyllae densities and feeding intensities were similar onS.phlebophyllaandS.reticulataand grazing intensity was comparable withP.palmenion S.pulchra. 5. A highly significant negative correlation was found between psyllid density and catkin dry weight in S.pulchra, S.phlebophylla and S.reticulata, suggesting that psyllid feeding is affecting catkin growth. 6. Predation of psyllid nymphs by syrphid larvae was heavy but there was no evidence of parasitism. 7. The life history strategies of the five psyllid species are discussed within the context of the constraints imposed by the arctic environment. This work forms part of the integrated programme 'Research on Arctic Tundra Environments' (RATE) designed to quantify the intensity and impact of herbivore grazing on tundra vegetation. The 'Rate' study area is investigate the distribution, abundance and phenology of the jumping plant lice (Psylloidea) associated with dwarf willows (Salix species) and t o measure their impact on their host plant. The Psylloidea is a group of small, host specific, phloem feeding insects which often occur at very high densities on dicotyledenous plants, particularly in arctic or alpine situations (Hodkinson, 1973a, unpublished; Koponen, 1973 Koponen, , 1978 Koponen & Ojala, 1974) . The Alaskan willows are particularly diverse; Hultdn (1 968) records fifty-six species for the region. Similarly the associated psy llids exhibit parallel diversification; twentyone species are recorded by Hodkinson (1 978) . At Meade River five psyllid species are associated with nine Salix species: their host plant relationships are summarized in Table 1 . Both Psylla palmeni Low and Psylla phlebophyllae Hodkinson were common and widespread, Psylla zaecevi Sulc and Trioza otkasookensis Hodkinson were widespread but uncommon, whereas PsyIIa macleani Hodkinson was found a t only one restricted locality. Hence attention was focused mainly on the t w o common species. Psylla palmeni and P.zaecevi are widely distributed throughout the far northern region of the Palaearctic and the latter species also occurs as a relic population in the mountains of northern India (Klimaszewski, 1973; Mathur, 1975) . Psylla phlebophyllae and T.atkasookensis are also recorded from the arctic regions o€ eastern Siberia (Hodkinson, unpublished) whereas P.tnacleani occurs only in Alaska (Hodkinson, 1978) . The Meade River area is topographically diverse with polygonized tundra, lake basin, dry ridge, bluff, sand dune and river edge habitats occurring in close proximity. The distribution of Salix reflects this habitat diversity and study sites were chosen accordingly. Salix pulchra is found mainly in the wetter more stable areas, in drained lake basins and around low centred polygons. Salix phlebophylla and S.reticulata occur together on the drier, more exposed ridges but the latter species also extends into the wetter lake basins. Salix lanata, S.alaxensis and S.glauca occur most commonly in disturbed situations, mainly along river margins. Both S.phlebophyIla and S.reticulata are chamaephytes, reaching a maximum height of about 4 cm, whereas the remaining species are phanerophytes reaching a maximum height of about 1 m, except S,alaxensis which reaches about 2 m in sheltered micro-habitats on the Dune site. Mine Site. Previous disturbance associated with mining activities had produced a diverse mixture of Salix spp. which made this site an ideal experimental area for the main comparative studies. The dominant vegetation was S.pulchra, S.lanata, S.alaxensis. S.glauca and S.reticuIata with less S.niphoclada. The 50 X 50 m sampling area was marked out as a 1 m2 grid. Dry Ridge, This was the main dry site containing Salix phlebophylla and S.reticulata growing in a mixed chamaephyte shrub community. The 10 x 1 0 m m sampling area along the ridge crest was marked out as a 1 mz grid. Lake Basin. This was the main wet site, subject to waterlogging when spring snow melted. Salix pulchra was dominant but S.reticu1ata wasalsoabundant. The 10 X lOOm sampling area was marked out as a 1 mz grid. Ice Scrape. This disturbed site along the edge of Meade River, was subject to ice scrape as the river ice melted. The willows were partially flattened and there was little or no ground vegetation. Salix alaxensis grew in a narrow strip along the edge of the river with a wider belt of S.lanafa behind it. The S.knafa extended into an undisturbed area with a welldeveloped ground flora. A sampling grid was not marked out. Sand Dunes. On this area Salix a h e n s i s grew in bare sand on a well-developed dune system adjacent to Meade River. Atkasook Village. A single colony of Psylla rnacleani was discovered on this site at the end of the 1976 season on S.alaxensis growing in a mixed Salix community adjacent to Meade River. No further studies were made on this site. A preliminary survey of the distribution of psyllid eggs on the willows at the various sites was carried out from 20 June to 3 July 1976. A minimum of nine catkin-bearing stems was removed from the plant and the number of eggs on the catkins ( d or 9). on the woody stem and on the green buds or leaves (depending on phenological state) was recorded. At the same time, a semiquantitative assessment of adult psyllid abundance on each Salix was made by taking ten sweep net samples (twenty sweeps per sample). Adult sampling was continued throughout the season to obtain material for investigation of ovary maturity. All data on female maturity are based on a minimum of fifty dissected females per host plant for each sample date. Subsequently, each Salix species on the Mine, Dry Ridge and Lake Basin sites was sampled for psyllids a t regular intervals during the summer of 1976. For each Salix species a minimum of twenty-four random samples was taken. Each sample was stratified to include three components: a female catkin, the current year's growth on the catkin bearing stem, and the current year's growth on a noncatkin bearing stem. The number of psyllid eggs and nymphs on each was then counted separately. Catkins of S.pulchra, S.Zanata, S.alaxensis and S.g&uca at the Mine Site were also sampled for P.palmeni nymphs on two occasions during July 1977. Preliminary sweep netting and egg sampling on 14 July 1976 suggested that psyllids were absent from the Ice Scrape site. To investigate the distribution of psyllids there a transect was made on 2 August 1976 from the riparian S.ahensis through to the unaffected S. lanata. At 10 m intervals along three separate transect lines a sweep net sample for adults (twenty sweeps) was taken together with a four catkin sample which was examined for nymphs. It also became apparent that P.phlebophyllae nymphs were absent from S.reticuhta growing on the wet Lake Basin although they were abundant on the same Salix species 6 0 m away on the Dry Ridge. Thus, to examine €he distribution of P.phlebophyllae along a wetdry gradient, eight catkins were taken at 10 m intervals (5 m up the slope) along a transect rising from the Lake Basin to the Dry Ridge on 3 August 1976. Similarly, preliminary egg sampling of S.aIaxensis at the Sand Dune site on 27 June 1976 suggested that psyllids were likewise absent from this site. Thus twenty-five random composite samples, as previously described, were taken on 3 1 July 1976. During the period when eggs were present all Salix samples were hand sorted. In later samples nymphs were extracted from catkins in modified Berlese-Tullgnn funnels. Straight heat extraction, as used successfully for other psyllid species (Hodkinson, 1973a; Hams, 1971) , was found to be inappropriate as direct heat caused older catkins to burst, trapping the nymphs. This problem was overcome by suspending the catkin over water in a polystyrene cup using a pin driven into the cup base: the water then served as the collection fluid. Extraction efficiency, estimated by counting the remaining dead psyllids on twenty catkins after extraction, was at least 96%. The phenological state of each Salix at each of the four main study sites was observed at regular intervals throughout the 1976 and 1977 seasons and recorded as dates of f i t occurrence of each stage of development. Seasonal development of the willows was rapid within the 2% month growing season (Fig. 1) . On the basis of phenology the species were divided into two groups; group 1 with catkins appearing before leaves (S.pulchra, S.lanata and S.alaxensis) and group 2 with catkins appearing after the leaf buds have broken (S.glauca, S.phlebophyl1a and S.reticulata). Group 1 species were the first to com-mence growth. On the Mine Site catkins of group 1 species appeared at least 14 days before group 2 species whereas at the Lake B a s h the difference was reduced to 10 days. On the Dry Ridge Site group 1 species were rare but catkin appearance in group 2 species was at least 9 days later than group 1 species at the Mine Site. Catkin appearance in group 1 species on the Ice Scrape Site was delayed by 4-5 days compared to the Mine Site, although development to maturity was more rapid. The period between'catkin appearance and catkin burst represents the period favourable for psyllid growth (see later). In group 1 species at the Mine Site this was 40-43 days and in group 2 species on the Dry Ridge 36 -4 1 days. Salk glauca (group 2 ) is atypical in that catkin burst is delayed and good estimates of time of ripening are difficult to obtain. At Meade River the temperature threshold for Salix growth was about 6OC. In 1977 the growing season began 10 days earlier than in 1976 and t h i s is reflected in the advanced phenology of the willows in that year (Fig. 1) . However, the actual development time of catkins t o maturity remained almost constant between years. The life history of P.palmeni on S.pulchra. S.lanata, S.alaxensis and S.glauca at the Mine Site and on S.pulchra at the Lake Basin during 1976 is shown in Fig. 2 . At snow melt in mid-June, prior to the commencement of Salix growth, sexually mature P.palmeni adults were present on the plant. Oviposition on catkins occurred as soon as they appeared (see later); the eggs hatched and the nymphs passed through five instan before the first new generation adults appeared in late July. These adults remained sexually immature at least until late August. It was not possible to investigate the mode of overwintering as the site was inaccessible during the winter. However, adults probably overwinter in the litter layer or in rolled dead leaves attached to the plant. The closely related Psylla moscouita Andr., which has a similar life history on Salix repens, hibernates in this manner (Hodkinson, unpublished) . The time between first appearance of eggs and fiirst appearance of new generation adults varied between 40 and 41 days. This period, making due allowance for observational error, is very close to the estimated life of the female catkin (40-43 days in phenological group 1 willows). Therefore, if the psyllid is to complete its life cycle within a female catkin (see later) there must be almost perfect phenological synchronization between the two. This is achieved by P.palmeni being ready to oviposit the moment the catkin appears. This strategy results in the vast majority of t g g s being laid o n the phenological group l willows pulchra. lanata and alaxensis. On the Mine Site S a l k grcluca and S.reticulata were 16 and 14 days later respectively in producing catkins by which time the majority of P.palmeni females were spent. There is some evidence, however, that a small number of eggs were laid on Sglauca and that these developed into adults late in the season. Salix glauca may thus act as a 'safety net' species on which P.palmeni can reproduce, albeit at a low level, in years when synchronization with group 1 species breaks down. S.phlebophylla at the Dry Ridge had a very similar life cycle to P.palmeni, except that eggs were not laid until the catkins emerged, 12 days later than Spulchra on the Mine Site and the first adults emerged 6 days later than P.palmeni (Fig. 2 ) . Again there was close synchrony between the life cycle and the life of a female catkin: the time taken to pass from first egg to first adult was 36 days compared with a catkin life of 39 and 41 days for S.phlobophyila and S.reticuIata respectively. The scattered data collected on these species suggest that all overwinter as adults and pass through a single generation each year. Trioza atkasookensis nymphs were always found in pit galls on leaves and never in catkins, whereas P.zaecevi nymphs were always found in catkins, often in association with P.paImeni. While n o data are available for nymphal development of P.macleani at Meade River, late instars nymphs were collected from female Xalaxensis catkins at Fairbanks (Hodkinson, 1978) . Table 2 summarizes the presence or absence of P.palmeni and P.phIebophyIlae on all sample sites. Absence is generally defined as there being no egg, nymph or adult found in any sample on a particular site throughout the study period. However, a few eggs and nymphs of P.palmeni were present on S.reti-C U~Q~Q at the Mine Site early in the season but were absent from a later sample. The data to + - S.lanata + - S.glauca + - S.retimlata -? - S.alaxcnsis + - Lake Basin S.reticulata - - S. pulchm + - S.rcticula fa + S.phlebophylla - + S.lanata - - S.alaxensls - - Dry Ridge - Ice Scrape Dunes -support these overall conclusions are presented in the following sections. Table 3 illustrates the distribution of P.palmeni and P.phlebophyllae eggs on the various willows at all sample sites. These data demonstrate that psyllid eggs are highly aggregated on male and female catkins and that very few are laid on the rest of the plant. This is not surprising on willows in which the catkin appears before the leaves and in which initially only the catkin is available to the gravid female for oviposition, but it also applies to those willows whose catkins appear after the leaves. Female psyllids must therefore actively seek catkins for oviposition. However, eggs were laid abundantly on both male and female catkins even though the former withered and dropped from the plant before the eggs hatched. This represents a major source of mortality in the populations of both P.palmeni and P.phlebophyllae. It was only on the female catkins that the psyllids were able to continue their development. Except for the odd end of the season sample, when numbers were small, the densities remained consistently very much greater on catkins than on the rest of the plant (see Table 4 ). Moreover, most of the nymphs found on the catkin shoot were associated with leaf axils right at the base of the catkin stalk. To test whether instar composition This index was used to compare instar composition on the catkin against those on the catkin shoot and non catkin shoot using a paired comparison 't' test. No significant differences between plant parts were demonstrated (Table 4) . 36 A 19 A 9 0 10 - 10 0 A 0 A 0 0 0 A 0 A 0 0 0 A 0 0 0 shook To test the null hypothesis that there is n o linear relationship between numbers of nymphs on catkins and numbers on catkin Fig. 3 shows the distribution of P.palrneni along the Ice Scrape transect. No nymphs or adults were found on either S.alaxensis or S.lanata within the area affected by ice scrape but both became common once the unaffected S.lanata was reached. This strongly suggests that P.palmeni is prevented from breeding on the Ice Scrape. This may be because Psylla palmeni overwinters as an adult and becomes active in spring as snow melt occurs. Overwintering adults will require protected hibernation sites in the ground vegetationllitter layer and these sites are absent from the Ice Scrape. Furthermore, ice movement over the host plant and surface flooding, at a time when adults are becoming active, will further minimize their chances of survival. A similar lack of suitable hibernation sites may explain the apparent absence of P.palmeni from S.alaxensis at the Sand Dune site during 1976 where any leaf litter falling to the ground in autumn is soon blown away. The success of P.palmeni on S.alaxensis at the Mine Site is probably because the Salix community is subject to little or no natural disturbance. The plants are old and large, probably the residual colonizers following the cessation of mining activity. In the more natural unstable habitats of S.alaxensis, P.palmeni has few chances of reproducing. Fig. 4 were absent from the Lake Basin but appeared as soon as the transition on to the Dry Ridge was reached. Flooding of the Lake Basin during spring snow melt ensures that the dwarf S.reticulata is completely inundated at the time when P.phlebophyllae adults would be becoming active and they appear unable to survive and reproduce in these conditions. In contrast P.palmeni on the Lake Basin appears able to move to its taller host plant S.pulchra and thereby largely avoid the effects of flooding. Thus, on a broad scale, a combination of naturally occurring physical events appears to be acting to prevent both P.palmeni and P.phlebophyllae from breeding in areas which at first sight, on the basis of host plant availability, might be considered suitable. The failure of psyllids to recolonize these sites once the disturbances have passed suggests that the rate of dispersal out of the more favourable areas is too low to overcome the host plant synchronization demands placed on the psyllid. The contrast between a late (1976) and an early (1977) growing season permits a comparison between the progression of psyllid development and host plant phenology. In the early season S.a&xensis was the first to commence growth (Fig. 1) and psyllid development on this species, as indicated by the index of advancement, was advanced over S.pulchra, S.1anata and S.g&uca (Fig. 5 ) . However, in the late season S.alaxensis commenced growth after S.&nata and Xpulchra (Fig. 1) and in consequence the development of P.palmeni nymphs on these species remained advanced relative to S.alaxensis (Fig. 5 ) . In both years psyllid development was latest on S.glauca but the difference was most accentuated in the late season. Catkin and shoot size varies considerably between different willows. For instance, the biomass ratios of catkins of S.lanuta, ahxensis, pulchra, glauca, reticulata and phlebophyllu were 15 : 15 : 5 : 5 : 1 : 1 respectively. Thus psyllid numbers per sample unit do not give a meaningful comparison of numbers per standard unit of habitat. Furthermore. it proved logistically impossible to sample all However, a few general conclusions can be drawn from a straight comparison of numbers (Table 4 ). On the Mine Site during 1976 numbers of P.palmeni nymphs per catkin were greater on S.pulchra than on S.lanat4 or S.alaxensis. but these three species as a group supported much larger populations than S.glauca or S.reticulara. The data for the latter were distorted by a large batch of eggs from a single catkin in the 23 July sample. The 1977 catkin data (Table 5 ) show an identical sequence of abundance on the different willows. Numbers of P.palmeni on S.pulchra were of a comparable order of magnitude on the Mine Site and the Lake Basin despite a slightly larger catkin size a t the latter. On the Dry Ridge numbers of P.phlebophyllae per catkin were approximately equal on both S.phlebophylla and S.reticulata. To provide a direct comparison of the potential feeding intensity by psyllids the biomass of psyllids per unit biomass of plants was calculated using psyllid biomass data (Table 6 ). For P.palrneni (Fig. 6 ) feeding intensity was consistently greatest on S.pulchra both on the Mine and Lake Basin sites, being approximately 3 and 1 6 times greater than on S.lanafa and S.alaxensis. respectively on comparable dates. The figures for S.reticulata and S.glauca were consistently low. For P.phlebophyllae ( Fig. 6) feeding intensity was similar on both For both P.palmeni and P.phlebophyllae psyllid biomass t o catkin biomass ratios were often less than 1: 100. The biomass figure at any one time represents a minimum figure for tissue production. In other psyllids the efficiency of conversion of phloem sap into body tissues is about 10% (Hodkinson, 1973a) . Thus t o have reached the observed biomass the psyllid population on the catkin must have consumed a weight of phloem sap at least one-tenth the weight of the catkin itself. Correlation analysis was used t o test the null hypothesis that psyllid density, measured as total number of psyllids per unit weight of catkin, was not related to the weight of the catkin. This analysis was carried out for each that there is an inverse relationship between psyllid density and weight of catkin, implying that psyllids are suppressing catkin growth, i.e. the larger the psyllid population the smaller the catkin. Significant relationships were found for those species already shown to be subjected to the highest grazing intensity. A parallel series of analyses was used to examine the relationship between psyllid density and weight of catkin bearing shoot. Significant negative comlations were found for P.palmeni on S.Lnata at the Mine Site on 22 July 1976 ( r = -0.40, P < 0.05) and Pphlebophyllae on S.phlebophylla at the Dry Ridge on 22 July 1976 ( r = -0.57, P < 0.01). These results, while not as clear cut as for the catkin, imply that the psyllids may be acting to suppress growth of the catkin bearing stem, at least in these species. The harsh arctic climate poses severe adaptational problems for insects (Downes, 1965) . MacLean (1 975a) lists low winter temperature, short growing season with associated low temperatures, low primary productivity and the weakness of photoperiod timing clues during the arctic summer as the main environmental variables to which invertebrates must adapt. He suggests that adaptation is achieved by a combination of life cycle adaptations, microhabitat selection, metabolic and timing adaptations together with resistance to winter cold. This provides a useful framework in which to discuss adaptation in P.palmeni and The basic resource of any sap feeding phytophagous insect is its host plant. In general growth and reproduction can only occur while the host plant itself is growing, although there is some evidence for aphids that feeding on plants can occur prior to bud burst (Hill, 1962) . Thus if the plant is growing for only a short period, a high degree of phenological synchrony between insect and host plant is essential if the resource is to be fully exploited. The life cycles of both P.palrneni and P.phlebophyllae a t Meade River show a high degree of synchrony with host plant phenology and this, coupled with a rapid development rate, enables them t o complete a generation within the summer season. Such a life history is essentially the same as in the equivalent temperate univoltine Salixfeeding species Psylla rnoscovita and Pmliceti Forster where growth and reproduction occurs during the short period of the year when conditions are favourable (Hodkinson & White, 1979) . This shows a marked contrast with the heather psyllid Strophingia ericae (Curtis), in which the evergreen host plant is 'available' throughout the year, and which has adapted t o more rigorous climatic regimes by extending its life cycle from I t o 2 years (Hodkinson, 1973a, b; Parkinson & Whittaker, 1975) . Several adaptational features permit such a rapid life cycle. Firstly, both P.palrneni and P.phlrbophyllae lie t o the left of the r-k selection spectrum as defined by Southwood (1977) and possess attributes, such as high reproductive/mortality rates, necessary for the rapid exploitation of a temporary resource, albeit a somewhat restricted food resource. Secondly, the psyllids select a highly favourable microenvironment on the plant, namely the female catkin. Johnson (pers. comm.) showed that levels of soluble nitrogen, the food source of psyllids, are greater in female catkins than in any other plant component growing on the Alaskan tundra. Furthermore, the temperature regime within a developing catkin is more favourable than either air temperature or the surface temperature of adjacent leaves. At Meade River the catkin itself has an effective 'life' equal t o or only a few days longer than the time required by the psyllids t o complete their development. Thus. successful completion 'of the life cycle demands a mechanism by which eggs are concentrated on the female catkin as soon as it appears. This can be achieved by the sexually mature females being present on the plant at time of catkin burst and seeking out catkins for oviposition. Ovary maturation in female psyllids prior to the Salix flushing suggests that they are able to obtain some nutrition from stems. Female psyllids d o not discriminate between male and female catkins and a significant proportion of eggs laid on the former are lost to the population when the catkin drops from t h e plant. Lack of photoperiod clues a t the time when psyllids emerge from hibernation suggests that synchronization with the host plant is temperature controlled. Natural selection should act t o ensure that the temperature threshold for psyllid egg maturation and oviposition is close t o the temperature threshold for Salix growth. Similarly if the psyllid population is t o survive and reproduce it must complete its life cycle within the 'life' of a female catkin. The relative development rates of psyllid and catkin, taken in the context of the day degrees available for a given site, may well set the limits of psyllid distribution. Winter snow cover on the Meade River site is thin and the willows are exposed t o extreme winter temperatures. To ensure survival the psyllids must seek favourable overwintering sites, most probably in the litter layer or perhaps in rolled dead leaves remaining attached t o the plant, and then return t o the plant in spring. It is only the active adult stage which is capable of such movement, a t least on t o the taller willows. It is generally postulated (e.g. Pianka, 1974) that arctic organisms have evolved broad, nonspecialized niche-habitat requirements and that this results in the low diversityhigh density situation observed among such organisms. If one considers the Salix community a t Meade River t o represent the total food resource of the psyllid community, then there is some evidence of resource partitioning, i.e. niche segregation, among the psyllid species. However, some flexibility is still retained. Psylla phlebophyllae occurs only on S.phlebophylla and S.reticulata whereas Psylla palmeni occurs only on S.pulchra. S.lanata, S.alaxensis, S.glauca and S.niphoclada. While there is some temporal and spatial separation of oviposition by females of each species there is sufficient overlap for oviposition on alterna-ing, often of the order of 1 g of psyllid per lOOg of catkin emphasizes the fact that catkins are nutrient sinks which continually accumulate photosynthate produced else where in the plant and that P.pulmeni and P.phlebophyl&e are feeding at the site where the supply of high quality nutrients is maximal. Predation by syrphid larvae caused considerable mortality in both P.pulmeni and P.phlebophyIlue populations whereas no evidence of parasitism was found. This may again be a result of synchronization demands; predators such as syrphids are opportunistic generalists, feeding on any available psyllid instar as well as alternative food resources and having little need to synchronize their life cycle with that of their prey. In contrast, psyllid parasites are usually both psyllid and instar specific (Hodkinson, 1974) and hostparasite synchronization is highly vulnerable when subjected to a fluctuating environment. tive willows to occur. That this does not happen suggests that both psyllids are able to discriminate between willows and are thus specific to their respective host plant series. Further resource subdivision occurs within a single host plant. On Salix knata, P.pulmeni nymphs feed on catkins whereas Triozu atkasookensis nymphs feed in pit galls on leaves. However, some overlap of food niches does occur, with P.zuecevi nymphs coexisting with P.pulmeni nymphs on the catkins on S.lunutu and S.pulchru and P.mucleani occurring with P.pulmeni on S.a&xensis. albeit at different sites. With the exception of P.mucleani, a scarce species, all species exhibit a degree of polyphagy. This flexibility in choice of food plant may act as a mechanism for spreading the risk and coping with year-by-year climatic variation. For instance the later developingS.gluucu could act as a 'safety net' species on which P.pulmeni could complete its development in years when synchrony with other hosts broke down. The psyllid population data suggest that certain willows support higher populations than others and could therefore be considered favoured hosts. Psylla palmeni population densities were consistently higher on S.pulchru than on other SuIix spp. In contrast, S.lanutu and S.ulaxensis appear to be the preferred hosts of P.zuecevi and P.mucIeuni respectively (Hodkinson, 1978; Klimaszewski, 1973; Loginova, 1967) . Thus potential niche segregating mechanisms which could act under conditions of intense competition appear to exist. Variations in tundra topography lead to spatial separation of the different Sulix/psyllid associations. Physical perturbation of the environment by ice movement, seasonal flooding or blown sand appears to prevent psyllids breeding in certain areas which are naturally colonized by the host plant. The heterogeneous patterns of psyllid distribution are probably further accentuated by phenological variations within il single Sulix species, related t o the aspect of the site in which it is growing. For S.pulchra. S.phlebophyIla and S.reticulatu. those species subjected to the highest feeding pressure, psyllid populations appear to be suppressing catkin growth. The high psyllid biomass which catkins are capable of support- Adaptations of insects in the arctic Sampling pear foliage for nymphs of the pear psylla using the Lkrlese-Tullgren funnel (Psylla pyricola: Hem. Hom. Ryllidae) Jr Hodkinson. 1.0. (1974) The biology of the PsyUoidea (Hornoptera): a review The psyliids (Hornoptera: RyUoidea) of Alaska Psylloidea. Handbook for phe Idenifflcotion of Brit&h Insects. 2(5A) Herbivorous invertebrates of the mountain birch at Kevo, Finnish Lapland Notes on herbivorous insects of the birch in Southern Greenland On the mesofauna of the field layer of three subarctic habitats The jumping plant lice or psyllids (Hornoptera: Psylloidea) of the Palaearctic Review on palaearctic psyllids of the genus PsyJrO Geoffr. (Homoptera: Rylloidea). associated with willows Ecological adaptations of tundra invertebrates Ecology of Tundra Invertebrates at Rudhoe Bay, Alaska. Ecological Investigutions of rhe 7bndm Biome in rhe Ptudhoe Bay Regfon. Alaska (cd Seasonal patterns of abundance of tundra arthropods near Barrow Ryllidae of the Indian Subcontinent. Indian Council of Agricultural Research A study of two physiological races of the heather psyllid Srrophlngia ericoe (Curtis) (Hornoptera :Rylloidea We thank the Royal Society and the National Science Foundation (Grant OPP-75 -12943 to University of Alaska) for financial support and M n E. Hodkinson for writing computer data analysis programs. Hill, C.P.