Different species of grasses and sedges are common plants in the tundra and dominant in many of the most widespread vegetation types. These species (collectively called graminoids) are also important food plants for same of the most abundant herbivores in the tundra, such as certain species of lemming, reindeer/caribou, muskox and goose, and many studies show that graminoids tolerate the often extremely high grazing pressure by these animals very well.

Introduction

Why then study these plants and their interaction with herbivores? We believe that studies of graminoids are of key importance for understanding trophic interactions in the tundra and thus tundra community processes. It is of great scientific interest to understand the key to their grazing tolerance, and how they can persist with such dominance in the tundra. We also ask questions about the genetic differentiation of populations of the species we are studying. Genetic differentiation will be related to the prevailing environmental factors at the different sites we visited during the expedition and their geological history, which connects this project with Theme B – Biodiversity in the Arctic Tundra.

Although the main focus of this project is on graminoids, we also participated in the different parts of the vegetation inventories and other common Theme A studies.

Previous results

This project builds upon results from the Swedish-Russian tundra expedition to Siberia, Tundra Ecology 1994, where population dynamics and genetic variation were studied in relation to prevailing environmental factors in three common Carex species. Great differences were found between 14 sites in flowering and vegetative ramet (shoot) production (i.e. vegetative reproduction through clonal growth) (Jónsdóttir et al. 1999, Jónsdóttir, Fagerström, Stenström and Augner, in prep.). Some of these differences could be related to latitude, while the strongest explanatory factor was lemming population phase with the highest flowering frequency at sites where lemming populations were increasing. Trypsine inhibitor concentration relative to soluble proteins (TI/SPP) in these plants, which it has been suggested functions as induced chemical defence against monogastric herbivores (Seldal et al. 1994), varied greatly between the sites visited (Högstedt and Seldal unpublished). There was a negative correlation between TI/SPP in vegetative shoots and the frequency of vegetative ramets.

Vegetative reproduction is of great importance in maintaining these populations. It provides successful genets (the product of a single seed) with potential longevity of up to thousands of years (Jónsdóttir et al. in press). In spite of that, the level of genetic variation was often high, and most of the variation found was contained within each population (Stenström, et al. in prep.).

Hypothesis

We propose three alternative but not exclusive hypotheses to explain the relationship found during the previous expedition between plant flowering and lemming population phase pattern:

1. Grazing by cyclic herbivores depletes internal plant resources during the peak phase and thus reduces the probability of flowering and vegetative reproduction (clonal growth) in the plants, both being important components of plant fitness (i.e. how well individuals do in a population). Long preformation time for flower buds in Carex (buds form at least one year ahead) and their dependence on the short growing season cause a time lag in the plant responses, which explains why flowering frequency first reaches its peak during the increase phase of the lemming cycle
2. To be able to flower, the plants have to pass a threshold in the sum temperature experienced that accumulates over some years (Tast and Kalela 1971, Laine and Henttonen 1983). This causes fluctuation in the frequency of flowering and internal plant resources. High flowering frequency signals high nutritious status of the food plants, which results in lemming population increase until the population is limited by the availability of nutritious food
3. As TI/SPP ratio in the plant tissue is induced by lemming grazing, the quality of the plant tissue as food for monogastric herbivores is reduced (Seldal et al. 1994). Reduced quality results in decreased grazing pressure, and hence the frequency of vegetative ramets rises again after a lag-period.

Are plants really out of the question?

Many previous studies on the effects of grazing on graminoids have measured the production of above ground parts and/or chemical composition as response variables and they all demonstrate the extreme capacity these plants have for compensating for lost tissue and nutrients. However, very few studies address how heavy grazing and the subsequent compensation affects the population dynamics of the plants and their internal resource dynamics and, ultimately, their fitness.

The causes of the relatively regular population cycles of lemmings and other microtines in northern areas have interested many biologists over the past hundred years, and they still do, and many alternative hypotheses have been proposed (see Stenseth and Ims 1993, and report by Krebs et al. and Danell et al.). One of the hypotheses is that food quantity is the basic cause and another that change in food quality causes the cycles, but these hypotheses have received limited support. However, most previous studies have only considered above ground phytomass and have not distinguished between the different developmental stages of the plants (see review in Stenseth and Ims 1993). The question remains then: have these studies been detailed enough to reject the food quality hypothesis?

The above ground biomass of plants is not a homogeneous component, but a complex matrix of different developmental stages of plants and plant parts which may have different functional and nutritional status and which probably significantly influences their quality as animal food. Indeed, young plants are better defended than older ones (Bråthen et al. unpublished). Many herbivores are extremely selective and can choose between the different plant parts and developmental stages of the ramets (see report by Agrell and Berteaux). Furthermore, in tundra graminoids and many other tundra plants, only 10-20% of the total living biomass is above ground and the different ramets may be interconnected by below-ground rhizomes. Therefore, to understand the herbivore’s interaction with its food plants a detailed study is required where all the following plant aspects are considered: the population dynamics, the internal resource dynamics (i.e. how resources are allocated to different ramets and functions) and stage/age specific TI/SPP levels. In addition a detailed study is also required of if and how the animals choose between the different developmental stages and plant parts. By doing this we believe that this project may throw some new light on the food quality hypothesis.

Is the genetic basis or the physiological condition of our study species similar among the sites visited?

The scientific approach of the expedition, to carry out comparisons between widely separated sites along geographical gradients, demands control of the identity of the organism’s studies beyond species identification. Therefore, we also study the genetic variation within and between populations of some of the study species. But a genetic study is far more than just control, it can throw light on the biodiversity and phylogeography of tundra plants.

Although two populations may be similar genetically, their physiological condition may differ depending on environmental differences. Analyses of stable isotopes in the plant tissue (¹³C, ¹⁵N, ¹⁸O, and Deuterium), can reveal differences in the physiology of the plants at the time of tissue production (differential isotope sequestration depending on water use efficiency, nutrient use efficiency and water source) and thus provide a physiological fingerprint.

Methods and preliminary results

Species and choice of study population

At each site visited we studied one or two populations of a t least one of the two different genera Carex and Eriophorum, with the exception of the two northernmost sites (north-east Bathurst Island and Ellef Ringnes Island), where they were not found. Carex was represented by a member of the C. bigelowii s.lat. species complex (C. bigelowii s.str. or C. consimilis) and/or C. aquatilis ssp. stans, while Eriophorum was represented by E. angustifolium (Nomenclature follows Porsild 1957).

We always chose to study populations of Carex and Eriophorum in mesic sedge meadows on level or somewhat sloping ground towards south or south-west, where at least one of our study species was dominant.

Ramet frequency measures

Within the sedge meadow, between ten and fourteen 50×50 cm plots were semi-randomly chosen for ramet frequency measurement and plant sampling. In each plot the frequency of different developmental stages of ramets was counted: young vegetative (estimates the frequency of vegetative reproduction), old vegetative (distinguished from the previous one by attached withered leaves from the previous year), current year flowering (relative estimate of allocation to sexual reproduction) and previous year flowering (the total above ground withered part, estimates previous year’s flowering). All grazed shoots were recorded and all turds within the plots were identified and counted. This data has not been completely analysed yet, but we recorded large variations between sites in relative ramet frequencies (Carex: 1-26% flowering, 6-29% young vegetative; Eriophorum: 0-3% flowering, 7-35% young vegetative) and the frequency of grazed shoots.

Plant and soil sampling, vegetation cover

The cover of all vascular species and the total cover of mosses and lichens were estimated in each plot in the sedge meadow. In plots where mosses and lichens had more than 10% cover their depth was also measured.

At least one rhizome system of each species, with two or more interconnected ramet generations, was excavated from each plot and pressed dry. They will be used for more detailed analyses of ramet population dynamics, and subsequently analysed for the nitrogen and phosphorus content of the different ramet categories and plant parts. A part of this material will also be analysed for stable isotope content in collaboration with Jeff Welker, Wyoming University, Laramie, USA.

Another set of plants was excavated from the plots and quickly frozen in dry ice after dissection of the plants in to shoot and rhizome of young vegetative, old vegetative and flowering ramets. The shoots will be analysed for TI/SPP in collaboration with Göran Högstedt and the rhizomes will be analysed for storage of carbohydrates and nutrients.

Soil samples were taken for pH analyses, total nutrient content (N and P), total organic content and stable isotope analyses. Soil pH varied greatly demonstrating differences in the physical environment between sites.

Phenological controls and productivity index

Plants show large phenological changes in chemical composition over the growing season. Because the different sites could not be visited when the plants were at the same phenological stage, we need to make corrections for the phenological difference in our comparisons. We carried out two types of phenological control. First, fifteen clones of Carex bigelowii and Eriophorum angustifolium were collected just east of Iqaluit, Nunavut, at the beginning of the expedition, 26 June, and transplanted to a ”garden” on board the ship. They were harvested at intervals and treated in the same manner as described above. As a second control we quantified both flowering and vegetative phenological stages in all our plots. In estimating the vegetative phenological stage we took advantage of the withered leaves from previous year still attached to old vegetative shoots and measured the ratio between the longest current green leaf and the longest leaf from the previous year.  This shows that by the time we reached Wollaston Peninsula on Victoria Island (23 July) Carex had reached the peak of leaf growth, but Eriophorum reached that peak somewhat later. The length of the longest leaf from the previous year can also be used as a productivity index for the different sites. This index shows that productivity varies greatly between sites and that it is not simply related to latitude.

Simulated grazing experiment

To study the ability of the plants to increase TI/SPP in response to grazing at the different sites, a simulated grazing experiment was carried out. At all but the two northernmost sites, a population of Eriophorum angustifolium was located. A total of 40 ramets at least 1 m apart were marked. Half of these were subjected to simulated grazing, where approximately 30% of the leaf material was torn off. After 27 hours the remaining leaves of all the ”grazed” and the non-grazed plants were collected and immediately frozen for later analyses.

Food choice experiments

To study whether our potential findings of differences in food quality between Carex ramet types, i.e. old vegetative, young vegetative and flowering, affect lemming food preference we conducted food choice experiments with the brown lemming in collaboration with Jep Agrell and Dominique Berteaux. The flowering ramets were clearly preferred to both old and young vegetative ramets, but how these preferences are related to the TI/SPP in the ramets remains to be seen.

Genetic sampling

At each site the Carex populations in the above studies were sampled at 3 m intervals on two parallel 45 m transects at 3 m distance from each other (a total of 30 samples per population). The sampling took place shortly before the camps were fetched by helicopters. Samples were stored in a cool box and transferred to a -80°C freezer on board. Samples will be analysed using isozyme and DNA techniques (RAPD) in collaboration with Hilde Nybom, Balsgård, Swedish University of Agricultural Sciences.

Data analyses

The data on ramet frequency, the chemical composition of different ramet categories and plant parts will be compared between sites and analysed in relation to abiotic (climate, temperature sums, pH) and biotic (herbivore) factors. For estimates of herbivore populations we rely on data collected by other projects of Theme A (lemming population densities and lemming population phases, Krebs et al. this volume, Danell et al.), and data collected collectively by Theme A members (estimates of other herbivore densities based on turd counts). Unfortunately lemming populations were at low levels at most of the sites. Our data will be available for overall synthesis of the theme on trophic interaction.