In the Arctic, willows are often dominant in the terrestrial vegetation. They play important ecological roles by being food plants for many herbivores, adding organic material to the soil and stabilizing it through their extensive root systems. There are relatively few species of willow in the High Arctic. Salix arctica, S. herbacea, S. polaris and S. reticulata extend furthest to the north. Salix arctica has widest distribution in the Nearctic, but its distribution in Eurasia is more limited than that of the other species mentioned. In the Low Arctic there are about 15 species and some of them like S. lanata and S. glauca have a very wide circumpolar distribution.

Arctic willows are long-lived and their ramets can be up to 20-50 years old, but the roots and stocks may be hundreds of years old. The annual diameter growth of the woody parts is clearly visible as growth rings and makes ageing possible. The width of the growth rings provides us with information on the growth conditions for every year. Furthermore, it is possible to measure the annual increments of shoot length. This means that willows constitute a rich historical archive supplying us with information on past climates.

The willows are an important food source for various invertebrate and vertebrate herbivores, e. g. sawflies, chrysomelid beetles, lepidopterans, musk-oxen, caribou, polar hares, lemmings and ptarmigan. Flowers and nectar are important for many imagines of dipterans, lepidopterans and hymeopterans. When the willows are fed upon during winter by e. g. lemmings and hares the bark is damaged and the annual growth rings make it possible to date when this damage occurred. By collecting samples of damaged willows we can obtain information on the feeding intensity of different herbivores in the past.

The willows are also well suited for other ecological studies. They are dioceses, i.e. individuals are either male or female, which means that the responses of the two sexes to various ecological factors can be compared. In addition to the sex differences, there is often a large variation in morphological characters, both between and within willow populations. One explanation for this variation is that hybridization and polyploidy are common among willows. However, isolation of Arctic populations could also give rise to genetic variation between populations. Isolation is an important precursor of evolutionary change.

Therefore, Arctic willows provide an opportunity for following speciation processes and studying how herbivores adapt to the ongoing evolution in their host plants.

Sawflies (Hymenoptera: Tenthredinidae) occur on willows up to the High Arctic and there is often a close relationship between the sawfly species and their willow hosts. Many sawflies are monophagus and highly specialized. For example, in many cases galling species prefer to lay their eggs on fast growing plant individuals and larvae growth and survival is highest on vigorous plants. When the plants grow older and senescent population densities decline and sometimes the sawfly population becomes extinct. We could expect this dependence to be even more pronounced in Arctic ecosystems where larvae must develop fast and the finding of good host plants is essential. The often close relationship between these herbivores and their food plants provides an excellent opportunity for studying preference patterns and the evolution of insect-host plant relations.

An important feature of insect-host plant relations is the secondary chemistry of the food plants, i.e. in the case of willows, their concentration of phenolic substances. In general, high concentrations of secondary chemicals reduce plant palatability for herbivores. However, highly specialist insects like sawflies are often attracted by high concentrations of these compounds instead. Inter-specific and intra-specific differences in plant secondary chemistry can therefore explain preference patterns and the distribution of insect herbivores.

Arctic systems are generally more simple than e.g. boreal forests. Nevertheless, our knowledge of the structure, dynamics and regulation of herbivorous insect populations and communities in the Arctic is very limited. A better understanding of these systems in general would also allow us to test general predictions regarding insect host preference, regulation and evolution.

Aims

During the Tundra Northwest 1999 our aims were:

1. to use Arctic willows as a historical archive for information on past growth conditions and densities of mammalian herbivores at the different sites
2. to study their sex ratio variation
3. to examine their secondary chemistry
4. to determine the genetic variation between potentially isolated populations of willows and insect herbivores in order to study evolutionary patterns in both the hosts and their host plants
5. to test ecological and evolutionary predictions on the oviposition behaviour of sawflies
6. to study the role of natural enemies in regulating populations of sawflies and lepidopterans.

Willows as archives on climate and herbivores

At all sites with Salix arctica we collected stems and shoots for measurement of the annual increments in length and width. We wanted to test whether the length of shoot formed annually would provide an estimate of the differences in summer climate between years, and also if they could be used as an integrated measurement of climate and soil conditions between sites.

At 12 sites we collected scars made by bark-eating mammals on willow stems to use as indices of the past population levels of these herbivores. Each scar will be aged and the frequency distribution of scar ages will hopefully tell us during which years there were e.g. lemming peaks at the sites.

We also collected ramets of S. arctica in order to analyse the intensity of browsing and its impact on the architecture of willow ramets. In the High Arctic most browsing was caused by lemmings. The die-back of ramets was most extensive at the most southerly sites where we noticed higher browsing pressure by other mammals (caribou and Arctic hare) and ptarmigan. We also noticed that browsing had a considerable impact on the branching architecture of surviving ramets. The structure of ramets can also be used in analysis of the population dynamics of lemmings. A high density of lemming population gradually leads to increased feeding on previous years’ shoots. Relative densities of shoot age classes in ramets correlated with the actual densities of browsers. It is easy to detect the time for the last lemming population crash. After such a crash the willow ramets have grown normally without browsing. This method is also good for checking the willow scar method explained above.

Sex ratio variation of willows

In almost all willow populations studied there were more females than males, i.e. a female-biased sex ratio. To test this quite striking pattern we suggest three hypotheses:

l. male willows are more preferred by herbivores

2. male willows are more susceptible to parasitic fungi

3. male willows have a lower growth rate than females and are thereby less competitive

The palatability of male and female willow clones for lemmings will be studied by J. Agrell and colleagues and will test the first hypothesis. At one of the sites it was possible to test the second hypothesis. However, we found no gender related difference in susceptibility to the parasitic fungus Melampsora sp., 15.4% of male and 19.0% of female clones were infected. At five sites it was possible to test the third hypothesis and in all cases male clones were significantly smaller than female ones.

Secondary chemistry and herbivores

Very little is known about the secondary chemistry (phenolic substances) of Arctic willows. However, secondary compounds generally reduce plant palatability for herbivores and it is often assumed that these substances have evolved as a response to herbivore activity. Variation in secondary chemistry within and between willow species can therefore potentially explain the distribution and preference patterns of insect herbivores, especially in highly specialist monophagus species like sawflies.

Willow species, their hosts and evolution

Our main study species were Salix arctica, S. glauca, S. polaris and S. reticulata, which are wide-spread and common in the Arctic. From the DNA-samples of willows collected and the main insect herbivores (sawflies and the lepidopterans Arctic woolly-bears, Gynaephora groendandica and G. rossii) we will be able to determine the level of genetic variation between both different willow populations and the associated insect herbivores. The evolution of the two Gynaephora species is not clear. G. rossii probably evolved in Eurasia, but the origin of G. groenlandica is more uncertain. We therefore collected larvae and cocoons for DNA-analyses at all sites where woolly-bears were present. Hopefully, this will shed some light on the evolution of the two lepidopteran species. Further, we hope that the DNA-analyses will provide us with some of the high points in the evolution of plant-insect systems in the Arctic in general.

Plant secondary compounds can also be used for ad dressing questions of speciation and evolution in willows. The composition of secondary metabolites, such as phenolic glucosides, is under genetic control and species specific. By comparing the composition of phenolic glycosides in different willow species it is possible to determine a chemotaxonomy for willows. This chemotaxonomy also has implications for the taxonomy and evolution of both willows and closely associated insect herbivores such as sawflies.

We have therefore collected leaf material from all willow species encountered on this expedition. Later, we will determine the concentration of specific phenolic glucosides in the leaf material collected. This data will be used to enhance our understanding of the evolution and speciation of Arctic willow species and the evolution of their insect herbivores.

Earlier studies have shown that galling sawflies prefer to oviposit on fast growing shoots on young, fast growing ramets. The evolutionary explanation for this behaviour is most probably that larva growth and survival is greatest on long shoots. This pattern seems to be less pronounced for leaf-galling sawflies of the genus Pontania than for stem, petiole and bud galling sawflies of the genus Euura.

Consistent with this hypothesis, our results show that female sawflies preferred to oviposit in long shoots. The mean length of shoot with leaf-galls was significantly longer than for shoots without galls within the same ramet. This pattern was clear for both Pontania reticulata on S. reticulata  as well as for Pontania nivalis on S. glauca. There was also as light indication of a higher average shoot length on ramets with galls compared to ramets without galls. More detailed analyses of this data material will reveal if there is a strong positive relation between the oviposition preference of the sawfly females and the performance of the larvae. From an evolutionary perspective such a correlation would indicate that natural selection favours sawfly females that oviposit in long vigorous shoots.

Insect herbivore populations

The dominant herbivorous insect at some of the sites was the Arctic woolly-bear (G. groenlandica and G. rossii), a lepidopteran belonging to the Lymantriidae family. Larvae and cocoons of woolly-bears were found at most sites, but at site 11, Cape Bathurst they reached extremely high densities, up to 26.4 larvae/m^­2­or 264 000 larvae/ha. The densities of these polyphagous larvae, which prefer to feed on willows, Dryas, Oxyria and Eriophorum, was slightly higher in the drier Dryas habitats (26.4 larvae/m^­2) than in the more mesic habitats (18.8 larvae/m^­2). However, at most sites the densities were much lower. Despite the exceptionally high densities at one of the sites it is unlikely that these larvae are used as a food source by vertebrate predators. Hairy larvae of the families Lymantriidae and Arctiidae are usually not eaten by birds; the cuckoo is the only bird that feeds on them, and it is not to be found in the Arctic. However, many larvae were parasitized And seven different parasite species were reared out of the larvae during the expedition. This suggests that parasites might play an important role in regulating the populations of these species. This will be studied in further detail.

Conclusions

The Tundra Northwest 1999 thus offered many opportunities for studying ecological and evolutionary questions. Willows have the great advantage of being relatively species rich and common. They were easy to find at all sites, except on Ellef Ringnes Island, and were thus excellent study objects. These woody plants with their easily recognizable annual growth provided us with a rich historical archive on past climates and herbivore population densities. Female willows were more common than their male counterparts, and their greater growth rate suggests that they are better competitors than males. This is quite striking due to the fact that females invest more in reproduction than males. In addition, analyses of DNA and phenolic compounds will enable us to study the evolution of some willow species and their host plants in Arctic environments. It was also possible to study the oviposition behaviour of sawflies and the preliminary data analysed so far are very promising.