Search

We observed colonization and mortality of plants in small (26 m2) artificial soil disturbances and undisturbed controls in a tall-grass prairie from 1977-1980. This paper examines the dynamics of colonists based on mode of origin (seedlings and shoots) and broad taxonomic affinity (forbs and graminoids). Few seedlings colonized either disturbances or controls the first year, probably due to a severe drought in 1976 when few seeds were produced. Graminoid shoots were more abundant in disturbed sites than in controls the first year, probably due to the sSedge effect" of cutting roots and rhizomes and stimulating new growth. For each year following the first, the number of current residents peaked early in the season and then declined, with the decline more precipitous for seedlings than for shoots. The population fluctuations of seedlings were very predictable, and separate years resembled each other. In contrast to seedling populations, graminoid shoot populations generally were accumulating throughout the study. Populations of forb shoots rose and fell, but the years did not resemble each other. Neither seedlings nor forb shoots showed a response to the new uncolonized soil resource of the artificial disturbances compared to controls, but graminoid shoots were more common on disturbed sites. These results suggest a nonequilibrium dynamic among colonizers of small-scale microsites.

Abundance (g/m2) and diaspore weight are positively correlated in seven species of perennial grasses that occur in prairies. The rare grasses ( 10.0 g/m2) have heavy dispersal units (2.23 to 2.80 mg). This result is the first reported differentiating trait between related rare and common organisms occurring in same habitat. Three hypotheses that explain this phenomenon are compared; the third most likely holds. First, rare grasses may be rare because their small seeds are less successful in establishment than those of common grasses. Second, if the persistence of small populations is marginal, rare grasses may devote less energy (or other currency) to seed production. Third, rare grasses may be colonizers of spatially and temporally rare microsites appropriate for growth and thus have seeds adapted for longer distance dispersal than those of common grasses. This last hypothesis suggests a new pathway for the evolution of weeds.

Seven grass species on an abundance continuum from common to sparse on a Missouri tallgrass prairie were examined for two diaspore traits (morphology and weight), for three dispersal traits (terminal velocity, and mean and maximum lateral movement propagules released in still air from a height of 1 m), and for two parental traits (culm biomass at maturity and culm height at dispersal). Dispersal ability, measured in the laboratory as the lateral movement of propagules in the upper 5% (right tail) of the distribution of distances, is inversely correlated (r, =-0.94) with biomass abundance. Under constant conditions, seeds of sparse species travel farther than seeds of common species. Dispersal ability is more closely related to a populational attribute (abundance) than to a more proximate characteristic (diaspore weight). Diaspore weight is closely related (r, = 0.97) to culm biomass of the parent, in apparently simple allometry, and more weakly related (r, = 0.71) to biomass abundance. Large parents make large seeds. The sparsest species, which is small in stature, has light (0.08 to 0.19 mg), flattened diaspores with low terminal velocities (82 to 87 cm sec-1) and achieves lateral movement (up to 460 mm) by tumbling and gliding. Species of intermediate abundance have intermediate stature and diaspore weight (1.0 to 1.8 mg), unornamented propagules, high terminal velocities (191 to 209 cm sec-1) and intermediate dispersal ability (up to 388 mm). The three most common species have heavy propagules (2.2 to 2.8 mg), ornamented with long hairs. Although the ornamentation slows fall (96 to 194 cm sec-1), it acts as a "guide parachute," orienting the propagule so that it falls nearly straight downward. The common species, which morphologically appear adapted to longer dispersal, achieve little lateral movement (up to 235 mm) under laboratory conditions. Terminal velocity is a misleading descriptor of dispersal because diaspores with similar terminal velocities have differing movement patterns. When diaspores of two species (one tall common species and one short sparse species) are released in still air from their natural height, differences in maximum fall distance observed under constant height of release are lost. When dispersal is observed in the field by means of sticky seed traps, the slopes of the dispersal curves (regression coefficient of ln-transformed number of seeds caught on distance from the clump) do not differ from common and sparse species, despite the fact that seeds of the common species are released from greater heights. The orderly differences in dispersal behavior among species due to subtleties of morphology observed under laboratory conditions are masked in

The number of flowering stems for 82 species on a transect 2 x 400 m was counted twice weekly during 1978 at Tucker Prairie, Callaway County, Missouri, USA, a tall-grass prairie remnant. Phenological curves (number of flowering stems vs. day of the year) are narrower (smaller standard deviations) for wind-pollinated species than for insect-pollinated species. Symmetry (g ) does not differ significantly for wind- and insect-pollinated species. The curves are largely either symmetric or begin abruptly and end gradually (right-skewed). Overlap, measured as the fraction of the area under the curves occupied by two or more species, does not differ significantly for wind- and insect-pollinated species, among 20 groups of five randomly drawn species. Dates of peak flow- ering are distributed randomly over the season (May to November 1978) and their distributions do not differ for wind- and insect-pollinated species. Dates of peak flowering are also randomly distrib- uted for grasses, legumes, and composites considered separately. Thus, although species with differ- ent pollination modes show different shapes for phonological curves, the species aggregated into the "community" do not have ensemble patterns of temporal dispersion or overlap and cannot be dis- tinguished from a random collection.

The effects of plant form and emergence time on size hierarchy formation in populations of two morphologically and genetically distinct varieties of peas (leafless and leafed) were studied. There were no significant differences in germinability between the two varieties, although leafless peas imbibed more rapidly than the leafed ones did. Monocultures of leafed and leafless peas were established at two densities: plants grown alone in small pots and plants grown at 576 m-2. Time emergence was noted, and plant shape, biomass and seed production were measured at two-week intervals for ten weeks. Seedlings emerged continually over an eight-day period, and two cohorts of seedlings were distinguished (seedlings emerging 6-7 days after planting, and seedlings emerging > 7 days after planting). Dominance and suppression were observed in the high-density populations, and early-emerging plants had less hierarchical biomass distributions than did late-emerging ones. Although leafless peas were larger and suffered less mortality than leafed ones did at identical densities, there were no differences in the degree of size inequality between the two genotypes (emergence cohorts pooled), or within emergence cohorts between genotypes. The degree of size inequality increased with time among dominant individuals and decreased with time among suppressed individuals. These results broadly support Weiner and Thomas's (1986) hypothesis that plant form may affect the extent but not the existence of competitive asymmetry in plant populations. THE PURSUIT of causal mechanisms that generate unequal size distributions ("size hierarchies") in plant monocultures has been a recurrent theme in plant population biology over the last thirty years (e.g., Koyama and Kira, 1956; Ford, 1975; Harper, 1977; Rabinowitz, 1979; Gates, 1982; Turner and Rabinowitz, 1983; Hara, 1984, 1986a, b; Weiner, 1985; Benjamin and Hardwick, 1986; Weiner and Thomas, 1986; Ellison, 1987a). Two hypothI Received for publication 25 April 1988; revision accepted 15 September 1988. We thank Dr. E. T. Gritton (Univ. of Wisconsin) for supplying us with the peas, Dr. G. Marx (Cornell Univ. Agric. Exp. Station) for advice on growing peas, Dr. A. R. Watkinson (Univ. of East Anglia) for providing us a copy of Dr. Butcher's dissertation, and Dr. P. Dixon (Savannah River Ecology Laboratory) for providing the Gini coefficient program. Beth Farnsworth and Cliff Cunningham assisted with planting. Debra Vam Vikites and Steve DiCamillo provided invaluable assistance in harvesting plants and/or data analysis. Jeannie Ngai maintained the germination experiment. Phil Dixon, Beth Farnsworth, Sana Gardescu, Laura Jackson, and Jake Weiner provided constructive criticisms of early drafts of this manuscript that resulted in a much-improved final product. Supported by McIntyre-Hatch-Stennis funds through contract NY(C)183419. 2 Current address: Organizacion para Estudios Tropicales, Estacion Biologica La Selva, Apartado Postal 676, 2050 San Pedro de Montes de Oca, Costa Rica. 3Deceased 18 August 1987. eses that explain the genesis of size hierarchies in dense monocultures have been proposed: the dominance and suppression hypothesis, and the growth rate hypothesis (reviewed in Turner and Rabinowitz, 1983; Weiner and Thomas, 1986). Briefly, if dominance and suppression are occurring, then size distributions of individuals in dense, competing stands should be more hierarchical (sensu Weiner and Solbrig, 1984) than size distributions of noncompeting individuals. If, on the other hand, intrinsic differences in plant growth rate alone generate observed size hierarchies, there should be no difference between the degree of inequality among individuals of equal size in competing and noncompeting stands. While these two hypotheses are often considered to be mutually exclusive because each has a different predicted outcome (e.g., Turner and Rabinowitz, 1983), it is likely that growth rate and competitive effects interact to produce observed variation in plant sizes (Benjamin and Hardwick, 1986; Ellison, 1987a). To date, of the dozens of studies that explicitly have examined hierarchy development in plant populations (reviewed in Benjamin and Hardwick, 1986; Ellison, 1986; Weiner and Thomas, 1986), only two have supported the growth rate hypothesis. Turner and Rabinowitz (1983) found no differences in the

Sparse species have chronically small local population sizes, even though they occur in several habitats over a wide geographic range. Greenhouse de Wit replacement series with seven species of sparse and common perennial grasses of tallgrass prairie were performed with seedlings and tiller fragments for 5, 10, and 15, mo. As younger and older seedlings, sparse grasses overyielded and were advantaged by the interaction with common grasses. The common grasses underyielded and were disadvantaged in mixture with sparse grasses. As tillers, the interaction was less antagonistic, and both common and sparse grasses either overyielded or were unaffected by the interaction. Seedlings of sparse species were largest when planted in low proportion, surrounded by individuals of a common grass. Because the sparse species are not disadvantaged by interactions with their common neighbors, their competitive abilities are not implicated as a cause of their local rarity. Rather, the good competitive abilities of these spar...

In even-aged monocultures, distributions of seedling mass shift from symmetric to right-skewed as plants grow. One explanation for the shift is that numerous small individuals and few large ones reveal underlying dominance and suppression caused by depletion of resources. A second explanation is that the asymmetry represents a shift from normal to lognormal mass distri- butions because of variance in exponential growth rate. As a consequence of the first mechanism but not the second, the presence of resource depletion should increase both the skewness and the variance of distributions of individual plant mass. Even-aged populations of the prairie grass Festuca paradoxa were grown in the greenhouse for 44 d with and without competition (dense flats or one individual to a container), and with and without augmented nutrients. At each harvest, plants grown densely were smaller than isolated plants, dem- onstrating that resource depletion occurred. Distributions of mass of isolated plants skewed first and remained more highly skewed throughout the experiment. At a given mean mass, the standard de- viation of mass for crowded and isolated plants did not differ. At a given mean mass, the skewness of mass for crowded plants was not greater than for isolated ones. The second explanation is sup- ported, and shifts in distribution shape can be attributed to growth alone, in the absence of dominance and suppression. Competition in this case did not promote skewing but actually retarded its appear- ance. Consistency of these results with previous findings is discussed.

Biomass abundance and distance from a random point to the nearest flowering culm were measured during two seasons for eight species of native perennial grasses on a continuum from locally rare to common in a Missouri tall prairie. Mean abundance from the most common to the most rare ranges over three orders of magnitude (nine octaves) from 73 to .25 g/m2. The means of the distance to a culm range over two orders of magnitude (four octaves) from 0.3 m to 5.0 m. The logarithm of distance to a culm is linearly correlated with the logarithm of abundance. Distance to an inflorescence is greater for sparse species than for common species.

THE degree to which plants growing together interfere with one another by competing for resources such as light, moisture and nutrients is difficult to determine by direct measurement of resource depletion1. Plant ecologists have therefore looked for indicators of interference in population structure2,3. One proposed measure of interference is the shift that occurs from the ‘normal’ distribution of seed weight to a ‘log-norma’ distribution of the weight of seedlings grown from those seeds in an even-aged monoculture. This shift presumably reflects the emergence of a few large and dominant individuals and the suppression of larger numbers of small individuals. Recently Ford4 and Newbould5 have proposed that distributions of plant weight in competing populations are actually bimodal, resulting from the presence of distinct populations of large individuals with positive relative growth rates and small individuals with nearly zero growth rates. Even-aged plantations of Picea sitchensis demonstrate such disjunct growth rates, and mortality is confined to the smaller individuals4. Size distributions in naturally thinning populations of Abies balsamea and Prunus pensylvanica are also bimodal6. We have studied the distribution of seedling weight in even-aged stands of Festuca paradoxa, a native North American prairie grass7, and report here that although the distribution is bimodal, the bimodality is more striking at lower densities, where interference is less severe. This result is the reverse of that expected if the extent of bimodality were a direct indication of the magnitude of interference. Therefore, although bimodality may be an eventual outcome of interference, it cannot be used as a measure of the strength of interaction between individuals.

Characteristics of dispersal and establishment for water-borne mangrove propagules, including period and pattern of floating, period of obligate dispersal, time to root firmly, longevity, and vigor, are estimated for six species in Panama (Laguncularia racemosa, Avicennia germinans, A. bicolor, Rhizophora mangle, R. harrisonii, and Pelliciera rhizophorae). Dispersal properties correlate with the spatial distribution of adults within the swamp. Genera whose adults are found on higher ground, on the landward edge of the intertidal zone, have small propagules that require a period of freedom from tidal inundation of approximately five days in order to establish firmly in the substrate. Genera whose adults are found on the seaward edge of the swamp, in deeper water, have large, heavy propagules. MANGROVES ARE WOODY PLANTS that grow on quiet seashores in the tropics and subtropics (Walter 1971). Swamps composed of mangroves form an extensive component of tropical vegetation. Because vegetative propagation is an insignificant aspect of mangroves' spread and colonization, a knowledge of the dispersal and establishment of their propagules is a prerequisite to understanding the habitat division that results in adult distributions. This paper presents field and laboratory observations on some aspects of dispersal in four genera of Panamanian mangroves. Mangrove species show a characteristic zonation within the swamp (Watson 1928, Davis 1940, West 1956, Schnell 1971, Walter 1971, Thom et al. 1975). Some species are found predominately on the lower ground or more seaward portions of the swamp; whereas other species are on the higher ground or more landward portions of the swamp. The mangrove species on both coasts of the Isthmus of Panama and their zonational tendencies are shown in table 1. These trees constitute the entire higher flora of the swamps, except for epiphytic orchids, bromeliads, and a fern. Most mangrove species share two traits: dispersal in water (van der Pijt 1972) and vivipary. Macnae (1968) describes this latter phenomenon: "The zygote, once formed, develops uninterruptedly through the embryo into the seedling without the intervention of any resting stage." Gill and Tomlinson (1969) also discuss ". . . the continuous development of the embryo without any dormant period. In the usual sense, there is therefore no true seed." Essentially, the dispersal unit is a "small tree," for which the term propagule is convenient. The continual growth of the embryo during dispersal is a unique aspect of mangrove establishment. NATURAL HISTORY Laguncularia racemosa.-Laguncularia is a monotypic genus in the Combretaceae. L. raceniosa ranges from Florida and the West Indies to Brazil, from northwestern Central America to Peru, and from Senegal to Angola (Exell 1958). Laguncularia propagules are small and flattenedobovoid-ellipsoid (lens-shaped) (table 2 and figs. 1 and 2). They are pea-green (Munsell 2.5 7/10) when they fall from the parent tree but turn mnudbrown (Munsell 2.5 3/4) in two days or so. The TABLE 1. Mangrove species in Panama. Zone Pacific coast Atlantic coast Family Seaward Rhizophora harrisonii Leechman R. mangle L. Rhizophoraceae (lower ground) (= R. brevistyla Salvoza) Pelliciera rhizophorae Triana & Theaceae Planchon Landward Laguncularia racemosa Gaertn. L. racemosa Gaertn. Combretaceae (higher ground) Avicennia germinans L. A. germinans L. Avicenniaceae A. bicolor Stanley (formerly Verbenaceae) 1 Present address: Division of Biological Science, University of Michigan, Ann Arbor, Michigan 48109, U.S.A. BIOTROPICA 10(1): 47-57 1978 47 This content downloaded from 157.55.39.220 on Fri, 02 Sep 2016 05:18:39 UTC All use subject to http://about.jstor.org/terms pericarp serves as a float and is not shed until the propagule is firmly rooted or nearly dead. The propagules "germinate" while dispersing, the radide pushing out a plug surrounded by the calyx and elongating. Dispersing propagules commonly bear roots. The stages by which they take root in the soil are illustrated in figure 3.

(1) In order to measure total annual seed rain for a native grassland in North America, fifty 9 cm diameter sticky traps were exposed to collect dispersing propagules for 26 weeks in 1978. (2) 19 700 seeds m-2 were trapped (per trap for ln-transformed data, x = 126, x s = 91, x + s = 173). This figure is much higher (by two orders of magnitude) than previous reports for non-grassland habitats, but is an underestimate because vegetation partially obscured the trap surfaces and because dispersal continued into the winter when the traps became snow-covered. (3) There were two peaks of dispersal activity: one in early summer for graminoid species and another in early autumn for both dicots (mostly Solidago spp. and Aster spp.) and warm season grasses. The number of seeds falling per week is weakly correlated with the strength of wind for the previous week. (4) Thirty taxa were identified to species, and two taxa were lumped, probably containing an additional nine species. The distribution of the number of seeds caught among species appears uniform for logarithmic classes (strongly right skewed for arithmetic classes). (5) The seed rain was spatially patchy, and this result may strongly influence subsequent seedling interactions. (6) The species composition of the seed rain qualitatively resembled the flowering community much more closely than it resembled the seed pool in the soil. There is a 71 % reduction of seed density from the rain to the soil pool.

Larvae of the mangrove borer, Peocilips rhizophorae Hopkins (Coleoptera: Scolytidae), attack the hypocotyl tissue of viviparous propagules of the red mangrove, Rhizophora harrisonii Leechman (Rhizophoraceae), on the Pacific coast of Panama. Production and growth of roots by infested and uninfested propagules were observed in fresh and salt (sea) water aquaria for 104 days. Infested propagules produce as many roots as uninfested propagules. The destructive effect of the beetles' mines is offset by an increased tendency for infested propagules to produce roots on the hypocotyl above the excavated regions.

Propagules of four genera (Avicennia, Laguncularia, Pelliciera and Rhizophora) of mangroves in Panama were planted in reciprocal gardens with solid stands of each genus. In most cases, mangrove seedlings grow and survive in the habitats of the adults of other species as well or better than they do in the habitats of adults of their own species. Seedlings do not shown superior growth (measured as five traits) in those sites where their adults customarily occur. I found little support for the idea that mangrove zonation is controlled by the species' differential tolerance to or preference for physiological conditions occurring along the tidal gradient. An hypothesis is proposed that mangrove zonation may be controlled by tidal sorting of the propagules according to size and by differential ability of propagules to establish in deep water. In Panama and other portions of the world, mangroves found in the seaward and deep water portions of swamps have larger, heavier propagules than do mangroves found in the landward and shallow water portions of the swamps. Selective pressures for efficient dispersal and establishment in a tidal environment may override selec- tion for habitat expansion on one hand or avoidance of competition and habitat specialization on the other hand. The seed morphology of upland ancestral forms may be a major evolutionary determinant of propagule and hence dispersal characteristics in the unrelated Panamanian mangroves. Physiologic specializa- tion and optimal growth in a restricted band within the swamp may be precluded by fluctuating conditions that vary annually in concert over the entire swamp.

SUMMARY (1) Cohorts of seedlings of Avicennia bicolor+A. germinans, Laguncularia racemosa, Pelliciera rhizophorae and Rhizophora mangle were studied for 1 year, the Avicennia and Pelliciera on the Pacific side of the Isthmus of Panama, and the Laguncularia and Rhizophora on the Caribbean side. The seedlings were in almost pure stands of the mangroves concerned. (2) There was an exponential decay of numbers in three species, the half-lives being 85, 338 and 604 days in Avicennia, Rhizophora and Pelliciera respectively. (3) The rate of decay of the Laguncularia cohort increased with time; 50% were lost in

The soil bank of buried viable seeds in October 1978 at Tucker Prairie in Missouri, USA, was assessed from 30 sod plugs (5.8 cm diam., 12.0 cm deep) spread in greenhouse trays for 145 d. 17.1 ? 8.7 (x ? s) seedlings per plug emerged, equivalent to 6470 seeds m-2, a figure higher than previous reports for prairies. Most (74%) seedlings emerged during the first month. Seventy-nine per cent of the seedlings survived to be identified to 28 species, 18 of which were dicotyledons, and 10, monocotyledons. Juncus tennis comprised 59% of the individuals. Forty-six per cent of the species produced 1 or 2 individuals. The density of seeds in the soil shows a 71% reduction from the density of seeds in the seed rain. The majority of common species in the seed rain (7 of 10) appear in the soil bank. In contrast, many of the common elements of the bank (7 of 11) are not detected in the rain. These latter species probably have long-lived seeds with pulsed input, and the former species have more transient presence in the soil. Thus, the seed rain resembles the soil bank more than the bank resembles the rain.