Factors affecting the seed-setting success of Dryas octopetala in front of Broggerbreen (Brogger Glacier) in the high Arctic, Ny-ilesund, Svalbard Naoya Wada To determine the factors restricting plant reproduction in front of a glacier, the gender expression and seed production of Dryas octopetala L. (Rosaceae) were observed, as well as the grazing pattern ,Of reindeer on flowers, near Broggerbreen (Brogger Glacier), which is near Ny-Alesund (78"55'N, 11 "56'E), Svalbard. Three hundred shoots with flowers and flower buds were randomly tagged in early July 1996. Between then and the end of flowering i n late July, 100 (33%) flowers and buds were grazed by reindeer. Out of the surviving flowers, 145 (76%) shoots had hermaphrodite flowers, while 45 (24%) shoots had male flowers without a developed gynoecium. Male flowers, which appeared later than hermaphrodite flowers in the population, were significantly smaller than hermaphrodite flowers in dry weight. In the hermaphrodite flowers, moreover, smaller flowers showed lower dry-weight allocation to the gynoecium as compared to larger flowers. During the observation, hermaphrodite flowers did not produce any developed seeds under a natural condition (0% seed-set). Cross-pollinated flowers showed 8% seed-set. On the other hand, flowers which were artificially warmed in small greenhouses during the flowering period showed 60% seed-set, regardless of cross-pollination or autodeposition of pollen from anthers to stigma (self-pollination). Thus, it was found that grazing, gender variation in relation to the length of the growing season and the flower size, and - in the flowering period - low temperature rather than pollinator limitation strongly affected the seed production of D. octopetala in the population studied. N. Wada, Dept. of Biosphere Science, F a d 8 of Science, Toyarna University, Toyarna 930-8555, Japan. Introduction Global climatic warming is expected to have a stronger impact on Arctic terrestrial ecosystems than on lower latitudinal ecosystems (Maxwell 1992). Climatic warming in the Arctic will promote deglaciation, exposing bare ground in front of glaciers. To understand ecological succes- sion in deglaciated lands, it is important to clarify how spores or seeds invade deglaciated locations and what factors limit the seed production of plants found around glaciers. Environmental conditions in the high Arctic are extremely severe for the growth and reproduction of tundra plants (Kelley & Weaver 1969; Bell & Bliss 1980). Cold environments may strongly restrict various kinds of temperature-dependent physiological activities affecting photosynthesis. fertilization and reproduction. Low temperature also affects the activities of insect pollinators (Kevan & Baker 1983; Totland 1994a), resulting in fewer opportunities for cross-pollination and therefore restricted gene flow. Thus, biotic and abiotic factors may greatly affect sexual reproduc- tion and seed production, especially in plants inhabiting areas in front of glaciers. This study examines the effects of cross-pollination and artificial warming on seed production. Its purpose was to clarify how these factors affect seed-setting success in a circumpolar plant species, Dryas octopetala L., in frpnt of Braggerbreen (Brogger Glacier), near Ny-Alesund, Svalbard. The inten- sity of reindeer's grazing on Dryas flowers was also examined to determine its influence on the reproduction of the plant. Reindeer frequently forage not only on grasses, sedges and lichens Wada 1999: Polar Research 18(2), 261-268 26 1 (Jefferies et al. 1992) but also on the flowers of D. octopetala, Saxijragu oppositifolia, S. nivalis and Silene acaulis (Wada, personal obs.). Gender expression of some plant species varies along an environmental gradient and among habitat conditions (e.g. Kudo 1997). Under stress- ful or resource-limited conditions for reproduc- tion, smaller-sized plants have a higher rate of success in reproducing through the male function than through the female function (Lovett Doust & Lovett Doust 1988; Klinkhamer & d e Jong 1993; Kudo 1993). In a preliminary survey, it was found that D. octopetala had two types of flowers: hermaphrodite and male flowers (Wada et al. 1999). The present study reports on the phenolo- gical and size differences between hermaphrodite and male flowers, as well as the gender variations within hermaphrodite flowers in relation to flower size. This study aims to clarify: I ) the number of flowers of D. octopetalu which are grazed or survive during the flowering period; 2) the effect of grazing animals on the spatial distribution of Dryas flowers; 3) the effect of flower size on gender expression or allocation to female function; 4) the effect of factors such as pollen limitation and low temperature during a flowering period on seed production. Finally, this paper discusses factors affecting seed-setting success from the onset of flowering to the fruiting stage. Materials and methods Study site The study site wac located ca. 1.5 km from the edge of Broggerbreen (Brogger Glacier) !nd ca. 0.5 km from Kongsfjorden, near Ny-Alesund (78'55'N, 1 ln56'E), Svalbard. The monthly aver- age temperatures of June, July and August 1996 were 1.9"C. 3.8"C and 3.5"C, respectively (data from Norwegian Polar Institute). In early July 1996, a 10 m x 50 m plot was established on a ridge. This plot was dominated by D. octopetala but included Salix polaris, Saxifraga oppositijolia, Silene acaulis, O q r i u d i g p a . Polygonum vivipar- ion and bryophytes. Plant material (Dtyus octopetala L . ) D. octopetalu is widely distributed in the low Arctic in Alaska (McGraw 1985) and Scandinavia and in the high Arctic in Svalbard, Norway, Greenland, and Russia (HultCn 1959; RGnning 1969, 1996). This plant forms a cushion on the ground, and has one flower per terminal shoot. However, the cushion form is not always well- defined, and it was difficult to determine indivi- dual plants in the plot used in this study. The bowl- shaped flowers show sun-tracking behaviour (flower heliotropism), resulting in high tempera- ture in the gynoecium as compared to ambient temperature (Kjellberg et al. 1982). This may be necessary for D. octopetulu to produce seeds in cold environments, suggesting a species sensitive to solar radiation and temperature in terms of its reproduction. Flowering phenology and grazing pattern In the plot, 300 shoots (271 flower buds and 29 flowers) of D. octopetala were randomly selected and tagged on 1 1 July 1996. They were observed every day from the onset of flowering to the end, and the phenology of hermaphrodite and male flowers was recorded, as was the number of grazed flowers. Because reindeer grazed on the Dryas flowers but not the stalks (peduncles), the grazed flowers could be recorded by the remaining flower stalks on the shoots. On 10 August, a 4 m x 4 m quadrat was established in the study plot. The quadrat was divided into 64 0.5 m x 0.5 m subquadrats. Grazed and non-grazed flowers and the leaf cover of D. octopetala were mapped in each subquadrat to clarify the effect of grazing on the spatial distribution of Dryas flowers. The spatial pattern of initial flowers and grazed flowers was analysed using Morishita's IS index (Morishita 1959), with the following formula: I & = n X x i ( x i - - I ) / N ( N - I ) , where n is the number of subquadrats of a given size (e.g. 64 0.5 m x 0.5 m quadrats), xi is the number of flowers in a ith subquadrat, and N is the total number of flowers (in the 4 m x 4 m quadrat). Id > 1 shows clumped distribution; I S = 1. ran- dom distribution; I6 < 1, uniform distribution. I S was calculated at a different subquadrat size; 0.25 m2, 1 m2 and 4 m2 (for details see Morishita 1959, 1962). The significance of I S > 1 or < 1 was tested by using F test ( d f = n - 1. K'), calculated I = I 262 Factors affecting the seed-setting success of Dnus octopetula in front of Broggerbreen with the following formula: F = [Z6 (N - 1) + tl - N] / (tz - 1). Gender wriatioti of Dryas flowers Twenty-four hermaphrodite flowers and 20 male flowers were randomly collected in the study plot in late July 1996. In each flower, the length ( L ) and width ( W ) of four petals per flower were measured. Individual petal area (PA) was calcu- lated with the following formula: PA = (L x W ) x TC x 0.25, where each petal was estimated as an ellipse shape. After that, each flower was dissected into petal, androecium, gynoecium (hermaphrodite flowen), calyx and receptacle and dried for 48 hours at 8O’C. The dry weight of these parts was measured and compared between hermaphrodite and male flowers. Furthermore, to clarify a variation of sex allocation based on dry weight within hermaphro- dite flowers, “femaleness” was calculated with the following formula: Femaleness = G W / (GW + AW), where G W is the dry weight of the gynoecium, and AW is that of the androecium (cf. Lloyd 1980). In this study, femaleness was used as an indicator of the reproductive effort of female function, and the relationship between femaleness and the flower size (dry weight) was analysed to determine whether femaleness is size-dependent. Effects of cross-pollination arid artificial nurmitig on seed production To determine whether seed production is restricted by pollen limitation or by low temperature, four treatments were conducted within the plot: 1 ) hand-pollination (HP). in which the gynoecium of Dryas flowers was supplied with pollen collected from the other five flowers ( n = 25 flowering shoots): 2) netting (N), in which a flowering shoot was covered with fine polyester netting (mesh size < 1 mm’; n = 15) to exclude any pollen-transport- ing insects; 3) warming (W), in which a flowering shoot was artificially warmed using a small greenhouse made of steel wire and vinyl sheeting ( n = 15); and 4) hand-pollination and warming (WHP, tz = 15). The fine polyester nets and small greenhouses were set up on mature flower buds of Dryus shoots from 16 to 20 July. In HP and HPW treatments, Dryas flower buds were emasculated on 20 July, and cross-pollination was conducted when those petals completely opened on 24 July. On 4 August, all nets and greenhouses were removed when all petals dropped off the flowers. Flower buds occurring near the greenhouses were tagged as controls ( n = 25). On 20 August. these shoots were collected, and the seed-set per flowering shoot was compared among treatments. Following Wookey, Parsons et al. (1993), “seed- setting” is defined as the number of fruiting shoots having one or more mature seeds with well- developed feathery styles beyond the extent of the sepals. divided by the number of flowering shoots in each treatment. To confirm the warming effect of the small greenhouses, a small temperature data-logger was set up (StowAway TidbiT, Onset Computer Corporation) inside and outside the greenhouse on the ground. This measured the temperature at two-minute intervals from 1 to 8 August 1997. Results and discussion Grazing pattern and flowering phenologj In the tagged Dtyas flowers, the onset of flowering occurred from early to late July. Out of 300 reproductive shoots, 10 (3.3%) buds were im- mature and did not bloom. One hundred (33.3%) flower buds and flowers were grazed by reindeer: 23 in the pre-flowering stage (flower bud), 73 in the flowering stage, and 4 in the post-flowering stage. Reindeer tended to forage on Dryas flowers on the 5th day (?SD 2 days), ranging from the 1st to the 1 1 th day ( n = 69), after their petals opened. Thus, 110 (36.7%) shoots failed to produce offspring before producing seeds and fruits. Grazing by reindeer had a strong impact on the spatial distribution of D t y s flowers. As shown in Figs. 1 and 2, Dryas flowers were initially distributed in small, clumped patches at 0.25 and I m2 scales. However, grazed flowers were more aggregated than initial flowers in both scales, indicating density-dependent predation on Dryas flowers. Because a high density of flowers results in a high probability of cross-pollination by attracting pollinators (Thomson et al. 1982; Sih & Baltus 1987; Totland 1994a), a density- dependent grazing pattern might not only directly Wada 1099: Polar Research 18(2), 261-268 263 c----l 0.5 m Fig. I . Spatial distribution of Dryus flowers (0 = grazed flowers, n = 30; 0 = non-grazed flowers, n = 110) in the 4 m x 4 m quadrat. Grey-shaded area = leaf cover of D p u s ocfopetulu; dotted area = gravel (bare ground). decrease the number of flowers potentially produc- ing seeds and fruits but also indirectly reduce a chance of out-crossing i n the remaining (non- grazed) flowers. Phenological and size differences between hermaphrodite and male flowers and gender variation within hermaphrodite flowers In the tagged flower buds and flowers, 190 shoots succeeded in flowering without being grazed. However, 45 (23.7%) flowers did not have any gynoecium; that is, they were male flowers. The number of male flowers versus hermaphrodite flowers tended to be slightly higher i n late flowering shoots than in early flowering shoots (Fig. 3). When 190 shoots were divided into two groups, early flowering shoots (defined as those which began to flower on or before 15 July) and late flowering shoots (after 15 July), the ratio of male flowers (22) to hermaphrodite flowers (90) was 0.24 i n early flowering shoots and 0.42 (231 55) in late flowering shoots. I n later flowering shoots or plants, the growing period may be insufficient for the completion of seed maturation, resulting in reduced seed production or low seed- set (Spira & Pollack 1986; Kudo 1991; Totland ...., 0 1 I I I I 0 1 2 3 4 Spatial scale (subquadrat size, m2) Fig. 2. Morishita’s I6 values for initial flowers (0) and grazed flowers (0) at the three spatial scales: 0.25 In’, 1 mz and 4 m’. The broken line shows I 8 = 1 (random distribution). ** = P < 0.01; * = P < 0.05. 1994b, 1997). The high ratio of male flowers to hermaphrodite flowers in later flowering shoots may reflect that reproduction through the female function is difficult for late flowering Dryas flowers due to the short growing period. This selection pressure might strongly act on Dryas flowers, resulting in a higher occurrence of male flowers in later flowering shoots. Philipp et al. (1990) observed that D. integrifolia also had hermaphrodite and male flowers, and male flowers appeared later than hermaphrodite f’lowers in a population in Greenland. However, because phe- nological differences vary among populations - 30 2 0 1 - lo !3 z 1- 0 10 15 20 25 30 35 Days from 30 June & Fig. 3. Flowering phenology of hermaphrodite flowers (0, n = 145) and male flowers (0. I I = 45). 264 Factors affecting the seed-setting success of D t y s octopetulu i n front of Broggerbreen 0.8 0 U 1 2 6 10 14 18 22 26 Flower size (dry weight, mg) Fig. 4. (a] Fiower size (dry weight) of male flowers aiid (b) hermaphrodite flowers, and ( c ) the frequency of malelall flowers (%) i n relation to the flower size. (Philipp et al. 1990), further information concern- ing several populations is required. Dry weight was significantly lower in male flowers (mean & SD: 9.6 & 2.5 mg, n = 20) than in hermaphrodite flowers (16.9 f 2.3 mg, n = 24, see Fig. 4). Male flowers had a smaller petal area and lighter petal and androecium weight than hermaphrodite flowers (Table I ) , suggesting lower investments to reproduction through male flowers. All flowers with a dry weight under 10nig were male shoots, while all flowers having a dry weight over 16 mg were hermaphrodite shoots. Flowers ranging from 10 mg to 16 mg in dry weight were composed of both male and hermaphrodite shoots (Fig. 4). According to the frequency distribution of dry weight. male flowers did not completely 0 m B 8 E o'21 0' 2.2 2.4 2.6 2.8 3 3.2 In flower weight (mg) F i g . 5. Relationship between flower size (In flower weight = X ) and femaleness ( Y ) i n hermaphrodite flowers of Dryus ocfo- p e t a h ( Y = 0.306% 0.345, r = 0.573, ) I = 24, P = 0.0028). segregate from hermaphrodite ones; rather, the gender sequentially changed from a male- to female- (hermaphrodite-) biased expression as flower size increased (Fig. 4). Moreover, dry- weight allocation to the female function (female- ness) vaned even within hermaphrodite flowers, and femaleness was positively correlated with flower size to a significant degree (Fig. 5). Thus, smaller flowers tended to allocate dry-matter investments into male organs (androecium) more than into female organs (gynoecium), as compared to larger flowers. With respect to size-dependence, this result supports a previous study by Kudo (1993) in Anemone debilis Fish. According to Molau (1993). D. octopetda has been recognized as a typical gynodioecious plant Tuble 1. Size differences between hermaphrodite (n = 24) and male flowers ( n = 20) of D p a s o c f o p m h (mean ? S D ) . The differences were analysed with Student's t-test. Hermdphrodi te Male f-value P Petal area (mm2)* 43.2 2 9.9 29.7 11.1 4.24 0.0001 Petal weight (mg) 5.9 & 1.6 3.7 * 1.2 5.33 <0.0001 Androecium weight (mg] 2.3 & 0.5 1.8 ? 0.6 3.66 0.0007 Gynoeciuni weight (mg) 2.6 & 0.9 - - - Calyx and receptacle 6.0 ? 1.0 4.2 k 1.1 5.92 <0.0001 (mg) Flower weight (mgj 16.9 k 2.3 9.6 k 2.5 8.74 <0.0001 *Individual (mean, petal area per flower. Wada 1999: Polar Research 18(2), 261-268 265 species, composed of a mixture of hermaphrodite and purely female shoots. However, such a gynodioecy type was not found in this study, but male shoots in addition to perfect hermaphrodite shoots on the same cushions (individuals) were frequently observed. This suggests an andromo- noecy or androdioecy of D. octopetala, as reported by Philipp et al. (1990) for D. integrifolia. Although further detailed studies are needed to clarify the variation of sex expression in D q a s species in relation to various environmental conditions, D. octopetala might have high pheno- typic and genetic variations for floral sex alloca- tion, as sh$wn in this study. In Ny-Alesund, Wookey, Parsons et al. (1993) and Wookey, Robinson et al. (1995) examined seed-sets of D. octopetala in various simulated environments. However, it was unclear whether a seed-set was obtained by the number of fruiting shoots divided by the number of flowering shoots minus the number of male flowering shoots. If these studies did not exclude the number of male flowering shoots, they would have underestimated the actual seed-set. It is important to examine the variations of the number of male flowers among experimental treatments and/or from year to year when assessing the differences of the seed-sets under various environmental conditions. Efects o j cross-pollination arid artificial warming on seed production The small greenhouses increased the mean tem- perature by ca. 1.3'C, maximum 8.7"C and minimum -1.4'C, as compared to the outside temperature from 1 to 8 August 1997 ( n = 5040 measuring points, Fig. 6). The results of pollina- tion and warming effects on seed-set are shown in Fig. 7. In this study, i n the year 1996, Dryas flowering shoots did not produce any seeds in the control shoots ( n = 25) or in 145 tagged hermaph- rodite flowering shoots. The cross-pollinated and net-covered shoots showed 8% and 6% seed-set, respectively, while the artificially warmed shoots showed 60% seed-set regardless of cross-pollina- tion or autodeposition (internal transfer of pollen from anthers to stigma without any external vector). This shows the self-compatitility of D. octopetala in a population in Ny-Alesund, as previously reported by Molau (1993) in a natural population in Latnjajaure (68"22'N, 18"29'E), in northernmost Sweden. The net-covered flowering shoots succeeded in producing a small portion of 0 1 I 1 I I I I I Q -2 I I 1 I 1 I I 118 218 318 418 518 618 718 818 Date (1997) F i g . 6. (a) Temperature ("C) inside (Folid line) and outside (dashed line) the small greenhouse, and (b) the difference ( A Temp.). seeds, probably because the fine net cover increased temperature slightly by reducing wind. Wookey, Parsons et al. (1993) also demonstrated that D. octopetala showed 60% seed-set by an artificial warming treatment using open-topped 80 I . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . W c., . . . . . . . . . . . . . . . . . . . . . . !i 40- z .............. 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C N HP W WHP Treatments F i g . 7. Effects of pollinator exclusion (netting, N J , cross- pollination (hand pollination, HP). artificial warming ( W ) . combination of cross-pollination and artificial warming ( W H P ) on seed-set (%)of D p a s flowers. C shows control. Numerals i n parentheses on the bars show number of samples. 266 Factors affecting the seed-setting success of Dryus octopefulu in front of Broggerbreen polythene tents: while there was no seed-set in the control in Ny-Alesund. The duration of artificial warming by small greenhouses was shorter (only in the flowering period) and the degree of increased temperature was smaller in this study than in that by Wookey, Parsons et al. (1993). However, a higher seed-set in the short-term artificially warmed shoots indicates that tempera- ture during a flowering period greatly affects seed- setting success. Furthermore, it suggests that temperature-dependent physiological processes such as pollen germination, pollen tube elongation, and fertilization may be strongly restricted i n a natural population of D. octopetula near Brag- gerbreen. Therefore, it is necessary to examine pollen activity and fertilization under various temperature conditions to elucidate temperature- dependent physiological traits. Concluding remarks This study demonstrates that grazing on flower buds and flowers, gender variation in relation to length of growing season and flower size, and (in the flowering period) low temperature rather than pollinator limitation appear to affect strongly the seed-setting success of D. octopetuh. In this study, the warming effects on sex allocation to the female function in hermaphrodite flowers, seed viability (germinability), seedling establishment, and survi- val were not examined. However, these are very important parameters to predict the population dynamics of D. octopetulu or its succession around or in front of glaciers under warmed conditions. In recent studies, Wookey, Robinson et al. (1995) and Welker et al. (1997) reported that the phenological development of D. octoperdu is accelerated under experimentally warmed conditions. This finding corresponds with a lengthening of the growing season in autumn, greater degrees of seed-set. and a higher potential of colonization of the bare ground due to heavier seed weight and then higher germinability. For this reason. long-term studies are necessary which include manipulated experi- ments and the monitoring of the growth and reproduction of many tundra plants in the field. Taking into consideration both abiotic and biotic factors, such as population dynamics of insect pollinators and grazing animals, such studies would serve to clarify the changes of Arctic plant populations and ecosystems under global climatic warming. Acknowledgements. - I am very grateful to the staff at the Norwegian Polar Institute, KBKC and NIPR for their support during my research in Ny-Alesund. I also thank H. Kanda and G. Kudo for their encouragement throughout this study, and Y. Bekku and T. Nakatsubo for their assistance in the field work. This study was supported by a Japanese Ministry of Education, Science and Culture Grant-in-Aid for International Scientific Research (No. 06044224), Project “Ecosystem Change at the Glacier-Edge Areas i n the Arctic.” References Bell, K. L. & Bliss, L. C. 1980: Plant reproduction in a high Arctic environment. Arctic A l p . Res. 1 2 , 1-10, Hultkn, E. 1959: Studies on the genus D q ~ a s . Sverisk Bot. Tidskr. 53, 507-547. Jefferies, R. L., Svoboda, J., Henry, G., Raillard, M. & Ruess, R. 1992: Tundra grazing systems and climatic change. In F. S . Chapin I11 et al. (eds.): Arctic ecosystems in u changing climuret an ecophysiological perspective. Pp. 3 9 4 1 2 . California: Academic Press. Kelley. J . J . & Weaver, D. F. 1969: Physical processes at the surface of the Arctic tundra. Arctic 22, 4 2 5 4 3 7 . Kevan. P. G. &Baker, H. G. 1983: Insects as flower visitors and pollinators. Annu. Rev. Entoni. 28, 4 0 7 4 5 3 . Kjellberg, B., Karlsson. S. & Kerstensson, I. 1982: Effects of heliotropic movements of flowers of D n a s octopetuln L. on gynoecium temperature and seed development. Oecologiu 54. 10-13. Klinkhamer. P. G. L. & de Jong. T. J. 1993: Phenotypic gender i n plants: effects of plant size and environment on allocation to seeds and tlowers i n Cyiloglossum officinule. Oikos 67. 81- 86. Kudo, G. 1991: Effects of snow-free period on the phenology of alpine plants inhabiting snow patches. Arct. A / p , Rrs. 23. 43-43, Kudo. G. 1993: Size-dependent resource allocation pattern and gender variation of Anetnone dehilis Fish. flanr Species B i d . 8, 29-34. Kudo. G. 1997: Sex expression and fruit set of an andromo- noecious herh. feucedunum mu/tivittatwn (Umbellifirde) along a snowmelt gradient. Opera Bot. 32. 121-128. Lloyd, D. J. 1980: Sexual strategies in plants. 111. A quantitative method for describing the gender in plants. N . Z. J . Bot. 18, 103-1 08. Lovett Doust, J. & Lovett Doust, L. 1988: Sociobiology of plants: an emerging synthesis. I n J . Lovett Doust & L. Lovett Doust (eds.): Plant reproductive ecology Pattenis atid strategies. Pp. 5-29. New York: Oxford University Press. McGraw, J . B. 1985: Environmental ecology of D n u s octopetulci ecotypes. 111. Environmental factors and plant growth. Arct. A / p . Res. 17. 229-239. Maxwell. B. 1992: Arctic climate: potential for change under global warming. I n F. S . Chapin 111 et al. (eds.): Arcrir ecosptrtns it] LI chaiigitig dimcite: an ecopliysiologicul prrspectice. Pp. 11-34. California: Academic Press. Wada 1999: Polur Research 18(2), 261-268 261 Molau, U. 1993: Relationships between flowering phenology and life history strategies in tundra plants. Arct. Alp. Res. 25, 39 1-402. Morishita, M. 1959: Measuring of dispersion of individuals and analysis of the distributional patterns. Memoir Faculty of Srience Kyushu University Series E (Biology) 2, 215-235. Morishita, M. 1962: 18-index. a measure of dispersion of individuals. Res. l'opul. Ecol. 4 , 1-7. Philipp, M., Bocher, J., Mattsson, 0. & Woodell, S. R. J. 1990: A quantitative approach to the sexual reproductive biology and population structure in some Arctic flowering plant% Dr)us integrifolia, Silene acaulis and Ranunculus nivalis. Medd. Grgnland, Biosci. 34, 1-60. Ronning, 0. 1. 1969: Features of the ecology of some Spitzbergen plant communities. Arct. Alp. Res. 1, 29-44. Ronning, 0. I. 1996 Theflora of Svalbard. Pp. 66-67. Oslo: Norwegian Polar Institute. Sih, A. & Baltus, M.-S. 1987: Patch size, pollinator behavior, and pollinator limitation in catnip. Ecology 68, 1679-1690. Spira, T. P. & Pollack, 0. D. 1986: Comparative reproductive biology of alpine biennial and perennial gentians (Genriana: Gentianaceae) in California. Am. J . Bot. 73, 3 9 4 7 . Thomson, J. D., Maddison, W. P. & Plowright, R. C. 1982: Behavior of bumble bee pollinators of Aralia hispida Vent. (Araliaceae). Oecologia 54, 326-336. Totland, 0. 1994a: Influence of climate, time of day and season, and flower density on insect flower visitation in alpine Norway. Arct. Alp. Res. 26, 6 C 7 1 . Totland, 0. 1994b: Intraseasonal variation in pollination intensity and seed-set in an alpine population of Rununculus acris in southwestern Norway. Ecography 17, 159-165. Totland, 0. 1997: Effects of flowering time and temperature on growth and reproduction in Leontodon uuturnnalis var. raruxaci, a late-flowering alpine plant. Arcr. Alp. Res. 29, 285-290. Wada, N., Kudo, G. & Kojima, S. 1999: Gender variation of D n m octopetala along snowmelt and latitudinal gradients in the subarctic and the high Arctic. Polar Biosci. 12, 87-99. Welker, J. M., Molau, U., Parsons, A. N., Robinson, C. H. & Wookey, P. A. 1997: Responses of D v a s ocroperala to ITEX environmental manipulations: a synthesis with circumpolar comparisons. Global Change Biol. 3, 61-73. Wookey, P. A,, Parsons, A. N., Welker, J. M., Potter, J . A,, Callaghan, T. V., Lee, J. A. & Press, M. C. 1993: Comparative responses of phenology and reproductive development to simulated environmental change in sub- Arctic and high Arctic plants. Oikos 67, 490-502. Wookey, P. A,, Robinson, C. H., Parsons, A. N., Welker, J. M., Press, M. C., Callaghan, T. V. & Lee, I. A. 1995: Environmental constraints on the growth, photosynthesis and reproductive development of Dryas octopetala at a high Arctic polar semi-desert, Svalbard. Oecologia 1 0 2 , 4 7 8 4 8 9 . 268 Factors affecting the seed-setting success of Dryas octopetula in front of Braggerbreen