Winter Fruit Removal in Four Plant Species in Maine

We monitored individual plants of winterberry (Ilex verticil/ala), highbush cranberry (Viburnum opulus), wild lily-of-the-valley (Maianthemum canadense), and partridgeberry (Mitchel/a repens) during two reproductive seasons to measure the disappearance of fruits (presumably removed by fruit-eating animals) in plant species whose fruits ripen in early fall but persist through the winter. Patterns of fruit removal were variable between species, individuals, and years, but in general removal rates for all species were slow and relatively constant. Only Ilex fruits showed visible deterioration during the winter. Exclosures placed around Maianthemum plants in the field and preliminary feeding trials suggested that mice (Peromyscus sp.) and other small mammals may remove most fruits and disperse many viable seeds of Maianthemum, as well as other low-growing plants with persistent fruits, such as Mitchel/a. In the temperate zone the overwhelming majority of vertebrate-dispersed plant species ripen their fruits during the late summer and early autumn, a time which corresponds with the peak of fall migration of fruiteating birds (Burger 1987; Sherburne 1972; Skeate 1987; Stiles 1980; Willson 1986). As a consequence, in many species most fruits are removed over a short period (Sargent 1990). However, a small number of plant species retain their fruits through much of the late fall and winter (Borowicz and Stephenson 1985; Burger 1987; Jones and Wheelwright 1987; Sallabanks 1992; Stiles 1980). Such fruits can persist in the face of microbial attack and damage from freezing because of their low lipid content and the presence of secondary compounds (Cipollini 1993; Herrera 1982; Stiles 1980), but those same compounds presumably reduce the attractiveness of the fruits to birds and slow their removal rates (Herrera 1982; Jones and Wheelwright 1987; Sorensen 1983). Moreover, the diversity and abundance of active seed dispersers are relatively low during the winter; plants must rely on resident birds and mammals for seed dispersal, or maintain fruits until spring migrants reappear. The disadvantages of presenting fruits during the winter months are presumably outweighed by the benefits of avoiding the season of heaviest damage to fruits and seeds by microbes, fungi and insects (Herrera 1982), and minimizing competition with other fruiting plants for a limited number of avian seed dispersers (McKey 1975; Wheelwright 1985). In this study we examined disappearance rates of fruits of four wood* Department of Biology, Bowdoin College, Brunswick, ME 04011 (current address: Department of Wildlife, Humboldt State University, Arcata, CA 95521 (for reprints, write to N. T. W.)); ** Department of Biology, Bowdoin College, Brunswick, ME 04011. 16 Maine Naturalist Vol. 2, No.1 land plant species in southern Maine. The species -winterberry (/lex verticillata), highbush cranberry (Viburnum opulus), wild lily-:of-thevalley (Maianthemum canadense), and partridgeberry (Mitchella repens) -differ in habitat, secondary compounds, and fruit presentation patterns, but each produces persistent fruits eaten by vertebrates. Exclosures were used to determine the relative importance of birds, small rodents, and natural abscission in accounting for fruit disappearance in Maianthemum canadense. Preliminary feeding trials were conducted to clarify the role of mice (Peromyscus sp.) as potential seed dispersers or seed predators. METHODS AND STUDY SPECIES Our main study site was a mixed deciduous forest in North Yarmouth, Maine. Viburnum opulus was studied at a second site, in Brunswick, Maine, 20 km (12 mi) to the east. Censuses were conducted approximately weekly from late September, 1987 until low plants were covered by snow in December, and resumed at less frequent intervals from snow-melt until late April, 1988. Censuses resumed in early September, 1988 and continued at biweekly intervals until snowfall in mid-December. We refer to the September, 1987 to April, 1988 season as 1987, and the September, 1988 to December, 1988 season as 1988. Viburnum opulus L. (Caprifoliaceae), a shrub found along forest edges and in the understory, produces 50-500 bitter, bright red drupes on terminal panicles which persist through the winter (Sherburne 1972; Jones and Wheelwright 1987). The fruit, whose diameter averages 11.0 mm (SD [:t] = 0.5; N = 25), encases a single flattened seed (7.8 x 7.1 x 1.8 mm, N= 10). We found five shrubs of similar height (1.5-2.0 m) and fruit crop (300-400 fruits) growing along a dirt logging road in the forest. On each plant four infructescences, one from each side of the plant, were marked at the base of the panicle with a 1 x 5 cm piece of blue flagging. An average of 7.8 fruits were plucked from each infructescence to reduce it to 20 fruits and standardize infructescences (initial crops sizes of marked inflorescences in 1987: 27.8:t 7.9). Fruit crop sizes were smaller in fall 1988 (initial crops sizes: 17.6:t 5.5). Accordingly, in that period some infructescences were used with fewer than 20 fruits to repeat observations on the same shrubs (1987: N = 400 fruits; 1988: N = 313). Censuses began in both years before many fruits had been removed (empty pedicels per inflorescence: 1987: 1.3:t 1.8; 1988: 0.6:t 1.6). Ilex verticillata (L.) Gray (Aquifoliaceae), a dioeciouswetland shrub, grows to a height of 2.5 m and produces 10-1000 sessile orange-red drupes along its branches. The fruits, which are relatively dry, measure 7.7 mm in diameter (:t 0.5; N = 25), with .47 seeds 3.3-4.1 mm in length. Ten shrubs were haphazardly selected along a 200 m stretch of a swampy power line roughly 75 m wide, edged by mixed coniferous and deciduous 1994 J. A Gervais and N. T. Wheelwright 17 trees. On each plant, two branches bearing at least 20 fruits each were chosen. The number of fruits on each branch was reduced to 20, and branches were inconspicuously marked with twined grass at the base. In fall 1988, the same shrubs were censused (1987: N = 400 fruits; 1988: N = 411). Mitchella repens L. (Rubiaceae), a low-growing woody creeper of the forest floor, presents its red fruits singly in leaf axils. Its drupes, which have low water content, measure 7.8 x 6.4 mm (.:t 0.5; N= 6) and contain 4-7 seeds 2.9 mm in length (.:t 0.2; N = 20). Six plants growing beneath mixed hardwoods and conifers with little understory growth were haphazardly chosen and marked by blue flagging 1 m away. Plants were similar in size, covering about 0.5 m2 of ground, and each bore 2328 fruits (1987: N =157 fruits; 1988: N = 170). Once snow covered the plants in December, plants were not censused to avoid artificially exposing fruits to foraging animals. Maianthemum canadense Desf. (Liliaceae) is a clonal herb 4-15 cm in height (Worthen and Stiles 1986). Fruits are presented on a single raceme per plant (technically, a ramet or physiologically independent portion of the clone) and number 6-15 per infructescence. The mean fruit diameter was 5.9 mm (.:t 0.8; N = 12) with 1-2 seeds pe}:fruit averaging 3.5 mm in length (.:t 0.4; N = 16). In 1987, ten singly occurring fruiting ramets were located roughly 10 m apart along an old logging road in mixed forest. Each plant had 8-11 fruits (N= 89). In 1988,25 ramets (6-15 fruits each, N = 239) were monitored at the same site. Plant locations were marked by tying 8 cm of cotton string to sticks placed 20 cm away. Temporal patterns of disappearance of fruits as well as the firm attachment of fruits (periodically tested by shaking branches) and the general lack of fruits, seeds or pedicels beneath plants suggested that most fruits were removed by animals rather than simply abscised by plants. To determine the relative contributions of birds, small mammals, and natural abscission by the plant to fruit disappearance, in 1987 we constructed exclosures of 0.64 x 0.64 cm wire netting cut and folded into cubes 10 cm on a side. Ten exclosures had no access holes, and 10 had 4.5 cm square holes cut into each side but not the top. A large dense patch of fruiting Maianthemum was located and 30 plants haphazardly selected. The plants were divided into three fruit crop size categories (6-10, 11-13, and >13 fruits) then randomly spit into three groups. The first group's plants were covered by the complete exclosures (N = 86 fruits), the second group covered by the partial exclosures (N = 90 fruits), and the final group left as controls (N = 91 fruits). A mouse (Peromyscus sp.), captured in December, 1987 and maintained on an ad libitum diet of millet and sunflower seeds, was used in two feeding trials to explore the role of small mammals as seed dispersers of Maianthemum. Trials consisted of leaving only Maianthemum