Am. J. Bot. Li-Cor Advertisement
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (24)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Fort, K. P.
Right arrow Articles by Richards, J. H.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Fort, K. P.
Right arrow Articles by Richards, J. H.
Agricola
Right arrow Articles by Fort, K. P.
Right arrow Articles by Richards, J. H.
(American Journal of Botany. 1998;85:1722-1731.)
© 1998 Botanical Society of America, Inc.

Does seed dispersal limit initiation of primary succession in desert playas?1

Kevin P. Forta and James H. Richards2,a

a Department of Land, Air, and Water Resources, University of California, Davis, One Shields Avenue,Davis, California 95616-8627


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
To investigate the initiation of primary succession in a cold-desert playa-dune complex, we studied the large-scale (2000 m) seed (diaspore) dispersal patterns at Mono Lake, California. Seeds of seven of the ten species reaching the barren playa had wind-dispersal adaptations. Rates of dispersal (numbers of seeds per square metre per day) were as much as three orders of magnitude lower on the playa than in the diverse dune vegetation. However, seed input appeared sufficient to reach potential safe sites on the playa, with a peak input of 66 ± 8 total seeds·m·d. The smooth playa surface, the virtual absence of aboveground barriers, and the high windspeed environment promote the long-distance dispersal of seeds (at least 1300 m for Chrysothamnus spp. and at least 700 m for Sarcobatus vermiculatus). The large spatial scale of sampling revealed a relatively high seed input onto the playa by the dominant pioneer species S. vermiculatus, despite the low abundance of parent vegetation in this region. All of these results implicate low rates of seed entrapment as an obstacle to establishment on this desert playa, rather than a lack of seed input.

Key Words: Chrysothamnus spp. • desert • Mono Lake • primary succession • sand dunes • Sarcobatus vermiculatus; • seed dispersal • wind dispersal


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Seed dispersal plays a critical role in the spatial distribution and successional development of plant communities within many ecological systems. For example, exotic monocultures may persist by preventing propagule invasion from neighboring plants (Marlette and Anderson, 1986), colonizers of treefall gaps may originate from recent seed rain rather than the seed bank (Alvarez-Buylla and Martínez-Ramos, 1990), and efficient wind-dispersal mechanisms coupled with high seed production facilitates the relatively rapid and long-distance dispersal of some weedy species (Stallings et al., 1995). In primary succession where environmental conditions are not limiting, the content of the seed rain largely determines initial community composition and thus may set the course of subsequent community change. For example, the seed rain on deglaciated pioneer sites in Alaska contained little or no seeds of late-successional species (Chapin et al., 1994). However, when initial conditions are highly stressful, the order of seed arrival may not parallel that of establishment. Following the 1980 volcanic eruption of Mt. St. Helens, relatively slow-dispersed, stress-tolerant species became established first in some areas, facilitating the later establishment of more ruderal species present in the early seed rain (del Moral and Wood, 1993).

Desert playas of the western United States provide a novel arena for investigating the initiation of primary succession. Playas are basin landforms that lose water exclusively through evaporation and are subject to natural fluctuations in surface water levels at several time scales. These fluctuations expose large areas devoid of vegetation and propagules, making these areas available for primary colonization. At Mono Lake, California, lake elevation has varied by 40 m over the past 4000 yr (Stine, 1991). More recently, diversions of freshwater input streams for use by the city of Los Angeles have contributed to a lake-level decline of 14 m from 1941 to 1992, exposing ~81 km of lakebed (Los Angeles Department of Water and Power lake elevation database; Mono Basin Ecosystem Study Committee, 1987). Native and exotic plant species have colonized the exposed lakebed sediments very slowly or not at all. By 1991, plant cover over many square kilometres on the north shore was <0.01%. In some areas adjacent to the lakeshore, salt-encrusted alkaline soils clearly pose an insurmountable plant stress. However, large regions of the playa are not salt encrusted and soils there can support plant growth, as was demonstrated in a controlled-environment study using two dominant species from the local flora, Chrysothamnus nauseosus and Sarcobatus vermiculatus (Schaber, 1994).

Current understanding of seed dispersal surrounding desert saline lakes is highly limited; therefore, we first sought to describe the basic seasonal and spatial patterns by placing seed traps along a gradient ranging from unvegetated, recently exposed playa near the current shoreline to a relatively diverse desert shrub community on stabilized sand dunes ~2000 m from the lakeshore. Using these data, we could then investigate whether seed immigration plays a role in limiting plant recruitment on colonizable playa substrates. Specific hypotheses are: first, the total number of seeds reaching the nonvegetated playa is extremely limited, dropping to zero or near-zero over large distances (>10–100 m) from the nearest seed source. Second, a relatively low diversity of propagules arrive at barren playa sites due to the low diversity of proximal source plants, but increases in parallel to that of vegetation with increasing distance from the lakeshore.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
The study area was located on the north shore of Mono Lake, California, which lies at the eastern base of the Sierra Nevada mountains and at the western edge of the Great Basin desert (38°5' N, 118°58' W, elevation 1944–1966 m; Fig. 1). The shrubs Chrysothamnus nauseosus, Chrysothamnus viscidiflorus, Sarcobatus vermiculatus, and Tetradymia tetrameres dominated the plant cover on the stabilized dunes (>10% cover; maximum 31%), which comprised over 40 species (Table 1). Sarcobatus vermiculatus and Distichlis spicata dominated the region of depauperate vegetation between the diverse vegetation and playa, which was characterized by intermediate vegetative diversity and cover (1–10%). All units of dispersal analyzed in this paper were seeds or single-seeded fruits; the word "seed" was used synonymously for "diaspore." Hereafter we refer to S. vermiculatus as Sarcobatus, T. tetrameres as Tetradymia, Salsola tragus as Salsola, C. nauseosus and C. viscidiflorus, which have indistinguishable seed, as Chrysothamnus spp., and Eriogonum elatum and Eriogonum nummulare, which have indistinguishable seed, as Eriogonum spp.



View larger version (56K):
[in this window]
[in a new window]
  		Fig. 1. Location and map of study area. Compositions of diverse desert shrub vegetation, low-diversity depauperate vegetation, and vegetation on the playa are described in the text (see also Table 1 ). Dates of selected lakeshore berms that formed during declining lake levels are shown.

 

View this table:
[in this window]
[in a new window]
  		Table 1. Potential seed-producing plant species in the study area. Asterisks denote species that produced trapped seed in this study. Letters next to asterisks indicate species with indistinguishable seed. Species with uppercase dispersal type codes had distinct wind-dispersal adaptations on the seed (W = wing-like structures, P = plume, A = awn, H = long hairs); those with a lower case "s" were small seeded and lacked special wind-dispersal adaptations. Plant locations refer to study area zones delimited in Fig. 1: Dv = Diverse vegetation, Dp = Depauperate vegetation, Pl = Playa, SS = Sulfur Spring. All species are named according to Hickman (1993). Voucher specimens have been deposited in the University of California, Davis, herbarium (DAV).

 
The sparse vegetation on the playa (<1% cover) included only Sarcobatus and D. spicata and occurred almost exclusively on shoreline berms that were formed between periods of lake level decline (e.g., 1944 and 1957 shorelines, Fig. 1; Table 1). We observed no vegetation south of the 1200-m coordinate (i.e., the 1957 shoreline), with the exception of the 1984–1986 shoreline berm, which supported Sarcobatus, D. spicata, and five additional species. This berm is the only region on the playa within the study site that has continually recruited plants from the time of its formation to the present, based on the presence of seedlings and young, ring count and bud scar dated Sarcobatus at this location (J. H. Richards, personal observations). Seedlings and young plants are completely absent elsewhere on the playa (K. P. Fort, personal observation).

Seed traps were 4-L plastic pots, buried with the top edge flush with the soil surface. Each pot was 17 cm deep, with 16-cm diameter openings, and contained 4 cm of washed, 2-cm gravel rocks to discourage granivory and to prevent the redispersal of trapped seeds. Alternative trap designs permit the resuspension of seeds (Johnson and West, 1988) in order to integrate both the seed flux and entrapment components of dispersal. In this study, our primary consideration was seed flux, and the values obtained with our trap design are a measure of this component. However, because it is well established that most seeds do not travel far from the parent plant (Silvertown and Doust, 1993; Willson, 1993), our flux values may serve as an index of final seed distribution on the spatial scale at which we sampled. The precision of this index is expected to be lowest in areas with few barriers to dispersal, resulting in an overestimation of entrapment.

We installed the seed traps in shrub interspaces or on open playa on 12–13 September 1992. Traps were placed at 50-m intervals on each of eight 1.5-km-long transects, 100 m apart, aligned with the long axis of the study site (Fig. 1). The six northernmost traps on each transect were positioned 500 m west of the transect line to avoid the regionally atypical vegetation associated with a freshwater spring (Sulfur Spring in Fig. 1 and SS in Table 1). To assess seasonal variation, material falling into the traps was collected at the end of the following six sampling periods: 12–17 September 1992, 17 September–10 October 1992, 10 October–14 December 1992, 14 December 1992–14 April 1993, 14 April–23 June 1993, and 23 June–12 September 1993. Two additional collections were made for year-to-year comparisons: 12 October–12 November 1993 and 12 November–4 December 1993. Logistic constraints prevented the collection of every trap (total = 248 traps) during every sampling period; however, we collected the traps at 150-m intervals along the long axis of the study site (total = 88 traps) during all sampling periods. In winter months when traps were frozen into the ground, collections were made from five of the eight north-south transects (total = 55 traps). We placed additional traps (total = 32 traps) on and near the 1984–1986 shoreline berm (Fig. 1), which had several species growing on it in June 1993. These traps were sampled during the last three sampling periods (i.e., 23 June–4 December 1993).

We used sieving and hydropneumatic elutriation to separate soil and debris from the seeds (the latter method uses water jets and air bubbles to float organic material from a column of water to a screen; Gross and Renner, 1989), which were subsequently counted on a grid. Only Sarcobatus and Chrysothamnus were assessed for viability, using the criteria of an intact embryo. Previous work and experience propagating these shrubs from seed corroborate the use of an intact embryo as an indicator of >95% germinability (K. P. Fort, L. A. Donovan, and J. A. Schaber, personal observations). All analyses in this paper used counts of total seeds unless viable is specifically stated. Subsampling was used for samples containing large numbers of seeds. Seed input was expressed per square metre of trap opening and per day that the trap was open.

We sampled vegetation along eight 10-m radii centered on each seed trap on 18–19 August 1993 using the line intercept method. An anemometer (10 m height) 2 km from the study site logged wind run data at 1-h intervals (data provided by Great Basin Air Pollution Control District, Bishop, California). The dominant wind direction axis (N20° E–S200° W) was within ~5 compass degrees of the long axis of the study site, generally blowing onshore during the day and offshore at night (C. A. Toft, unpublished data). The mean wind speed was 10.7 km/h, with a maximum of 57.6 km/h. Of the total hours logged, 13.1% were >=20 km/h, classed as follows: 897 h with average wind speeds from 20 to 29.9 km/h, 317 h from 30 to 39.9 km/h, 93 h from 40 to 49.9 km/h, and 7 h from 50 to 57.6 km/h. Wind run on a per day basis is presented in Fig. 2A. For a general climatological characterization of the study site, see Toft (1995).



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 2. Wind run and seasonal abundance and viability of the seed rain for the three vegetatively dominant species. Viability was calculated after summing data from all traps within each sampling period. Viability was not determined for Tetradymia. Bar widths illustrate the length of the sampling period. Letters were generated from a polynomial contrast between the mean of the peak dispersal period, 17 September–10 October 1992, and those of the other five periods ending 12 September 1993; bars with the same letter are not statistically different. A log 10 plus 1 transformation was used to satisfy parametric assumptions. Abundance bars are backtransformed means; error bars equal one backtransformed standard error of the transformed mean, and are shown in one direction only when space was limiting. Note the change in scale following the break in the y -axis in (B).

 
Seasonal variation in seed abundance (numbers of seeds per square metre per day) for periods 1 through 6 (12 September 1992–12 September 1993) was analyzed for selected species with repeated-measures analysis of variance, using sampling period as the within-subject effect and following a log 10 plus 1 transformation to satisfy the assumption of homoscedasticity (SAS, 1988; Potvin and Lechowicz, 1990; von Ende, 1993). All such analyses excluded traps that were not collected during every sampling period. Because Mauchly's criterion indicated significant departures from compound symmetry in all cases, multivariate hypothesis tests were used. After significant differences were found between periods, a polynomial contrast was used to compare the mean of the peak dispersal period with those of the other five periods. Year-to-year comparisons of species-specific seed abundance were also made with repeated-measures analyses. For these, the high-dispersal mid- to late-fall sampling periods were used (October through December, using 10 October–14 December 1992 and pooling periods 7 and 8–12 October–4 December 1993).

The spatial patterns of seed dispersal were investigated using both a factorial analysis and a Cramér-von Mises-type test statistic (Syrjala, 1996). In the factorial analysis, a mixed model was used with zones (as depicted in Fig. 1) and species as fixed effects, and replicate north-south transects as a random effect (SAS, 1988). Post hoc mean separations were performed using the Ryan-Einot-Gabriel-Welsch Multiple F test. Unbroken transects were needed for the factorial analysis; therefore, the northernmost traps were assumed to lie 500 m east of their actual position (see Fig. 1). Although this adjustment was artificial, the results nevertheless agreed with the Cramér-von Mises-type test, which did not require the adjustment. This latter analysis tested for pairwise differences in the spatial distributions of selected species without any prior subdivision of the site into zones and transects. High values of {Psi} (Table 2B) indicate departure from the null hypothesis of no difference in the spatial distributions of the two populations being compared; the alternative hypothesis is an unspecified difference. This method was free of underlying theoretical distributions and removed the variability in abundance of the species by normalizing their distributions.


View this table:
[in this window]
[in a new window]
  		Table 2. Statistical analyses to characterize spatial patterns of seed dispersal. (A) Mixed-model factorial analysis and (B) Cramér-von Mises-type test statistic (see Materials and Methods) for pairwise comparisons of spatial distributions, using a Bonferroni adjustment to set {alpha} = 0.005. Analyses in both tables used only viable seed data for Sarcobatus and Chrysothamnus spp.

 
Least squares analysis was used for correlations of seed to vegetation abundance, seed rain to vegetation diversity, and seed abundance to wind run. N0 (species richness) and N1 (species diversity) were calculated according to Hill (1973), where N0 is the number of all species in a sample and N1 is the number of abundant species in a sample. Hill's N1 index is calculated as

N1 = e

where H is equal to Shannon's entropy. In practice, H must be estimated as

H = -{Sigma} [(ni/n)ln(ni/n)]

where ni equals the number of individuals of the ith species in the sample and where n equals the sum of all individuals in the sample (Ludwig and Reynolds, 1988).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Seasonal and spatial patterns of the seed rain
Seeds of 13 species (Table 1) were collected during six consecutive sampling periods (12 September 1992–12 September 1993) to compare seasonal abundance of seed dispersal. Five of these species comprised 98% of the total trapped seed: Chrysothamnus spp. (60.3%), Sarcobatus (17.0%), Eriogonum spp. (13.9%), Salsola (4.5%), and Tetradymia (2.7%). Each of the five dominant species showed significant differences in seed abundance during periods 1 through 6 (F5,51 >= 3.96, P <= 0.0041 for all multivariate repeated-measures analyses). Statistics for the remaining eight species could not be run due to insufficient sample size. Peak abundance occurred in the fall for all five species, with additional elevated seed abundance during the summer for Eriogonum spp. and Tetradymia (Fig. 2B–D illustrates these patterns for the three vegetatively dominant species). In addition to increased seed number, overall percentage viability increased for Chrysothamnus spp. and Sarcobatus during fall dispersal periods (Fig. 2B, C). Seasonal variation in seed abundance for each of the top five species did not correlate with wind runs summed within each sampling period (0.6437 < P < 0.9486).

Year-to-year comparisons revealed increases in seed abundance from 1992 to 1993 for viable Chrysothamnus spp. (F1,55 = 5.16, P = 0.03), viable and total Sarcobatus (F1,55 = 21.86–30.96, both P < 0.0001), and Tetradymia (F1,55 = 4.40, P = 0.04), but abundance declined for Salsola (F1,55 = 5.46, P = 0.02), and there was no difference for total Chrysothamnus spp. (F1,55 = 0.06, P = 0.80) and Eriogonum spp. (F1,55 = 0.71, P = 0.40).

The identification of ecologically meaningful spatial patterns was accomplished by only using data meeting three criteria. First, as in the analysis of seasonal abundance, only the dominant five seed-producing species were used. Second, only peak dispersal periods were used due to the large number of dispersing seeds and the highest percentage viability of Chrysothamnus spp. and Sarcobatus during these periods (Fig. 2B–D). Peak dispersal periods occurred during 17 September–10 October 1992 for Chrysothamnus spp., Eriogonum spp., Salsola, and Tetradymia, and during 12 November–4 December 1993 for Sarcobatus. Trapped seeds were also abundant for Eriogonum spp. and Tetradymia during 23 June–12 September 1993 and Sarcobatus during 17 September–10 October 1992, although only with Tetradymia was the abundance statistically indistinguishable from the peak period on a number of seeds per square metre per day basis (Fig. 2D). Analysis using these latter periods only (not presented) yielded similar results. Lastly, because the samples of Chrysothamnus spp. and Sarcobatus frequently contained large numbers of nonviable seeds, only counts of viable seeds from these species were used.

The factorial analysis of spatial patterns revealed no difference among replicate transects at any level of interaction (0.3729 < P < 0.9836; Table 2A). A significant species x zone interaction (P < 0.0001) was explained using post hoc mean separations, which showed higher seed abundance amid the diverse vegetation compared to the depauperate vegetation and playa for Chrysothamnus spp., Eriogonum spp., Salsola, and Tetradymia (mean separations not presented; however, Fig. 3A illustrates this pattern). Surprisingly, Sarcobatus showed a reverse trend wherein seed abundance was higher in the depauperate vegetation and playa compared to the diverse vegetation (Fig. 3C). This unusual pattern was corroborated by the Cramér-von Mises-type spatial analysis, which showed no difference between the spatial distributions of Chrysothamnus spp., Eriogonum spp., Salsola, and Tetradymia (all pairs analyzed individually), but distinguished the spatial distributions of each of the former species from that of Sarcobatus (Table 2B).



View larger version (23K):
[in this window]
[in a new window]
  		Fig. 3. Seed and vegetation distributions for selected species. Mean seed number was derived from eight replicate seed traps at each position along the long axis of the study site. Vegetation sampling is described in Materials and Methods. All error bars are ±1 SE. Note the change in scale following the break in the x -axis. Insets show correlations between seed rain abundance (total and viable) and vegetation magnitude, using the same units as corresponding measures from the main figures. Sampling periods for seed distributions were: (A) 17 September–10 October 1992; (C, E) 12 November–4 December 1993.

 
The contrasting patterns revealed in both spatial analyses were supported by correlations of seed abundance during peak dispersal to the cover of parent vegetation near each trap. Data for Sarcobatus did not correlate (total seed: r = 0.13, P = 0.20; viable seed: r = 0.05, P = 0.60) with this pattern due to the high abundance of seeds towards the lake where cover of the parent vegetation was low (Fig. 3C, D). In contrast, correlations using the remaining four dominant seed species, including both total and viable Chrysothamnus spp., were relatively strong (r = 0.56 - 0.82, P < 0.0001). Figure 3 (A, B) illustrates this relationship for Chrysothamnus spp.

Diversity and richness of the seed rain
During summer and fall dispersal periods, seed rain diversity remained approximately uniform along the north-south axis, in contrast to the decline in vegetation diversity (Fig. 4A). Correlations between seed rain diversity and vegetation diversity were significant but weak (r <= 0.45, P <= 0.04, except during 12 October–12 November 1993 when P = 0.29). Species richness was low on the playa in all sampling periods and in all vegetation types in late fall (October-December, data not presented). One or two species usually dominated the seed rain at any given location along the north-south axis, but the identity of the dominant seed varied along the axis (Fig. 4C).



View larger version (31K):
[in this window]
[in a new window]
  		Fig. 4. Diversity and richness of vegetation and seed rain. Vegetation sampling, species richness (N 0), and species diversity (N 1) are described and defined in Materials and Methods. Mean index value was derived from eight replicate seed traps at each position along the long axis of the study site. All error bars equal ±1 SE and vegetation x -coordinates in (A) and (B) are offset +10 m to prevent error bar overlap. Percentage seed composition was determined after summing seeds from the eight replicate seed traps.

 
Seed rain on the playa
Table 3 lists dispersal rates for the ten species that reached the playa. Chrysothamnus spp. and Sarcobatus contributed the largest number of seeds during each period. Seed input on the playa was as high as 190 seeds·m·d at a single trap during 12 November–4 December 1993, with a mean input of 66 ± 8 (mean ± 1 SE, N = 32) seeds·m·d during that dispersal period. At this time, viable seeds of the halophyte Sarcobatus dominated, accounting for 47% of the total trapped seed, and reached a maximum of 99 viable seeds·m·d at a single trap. Seeds from the eight minor species contributed between 0 and 9% of the total seed on the playa in any given sampling period. Seeds from Bassia hyssopifolia, a minor seed contributor, were recovered from various traps on the playa (Fig. 3E), yet this species occurred only on the 1984–1986 shoreline berm (K. P. Fort, personal observation).


View this table:
[in this window]
[in a new window]
  		Table 3. Seed rain on the playa from 1050 to 1500 m on the north–south axis of the study site (see Fig. 1). Values are mean seed number per square metre per day ± 1 SE. Atriplex canescens, Achnatherum hymenoides, and Muhlenbergia asperifolia were not detected on the playa during any sampling period. Annual total is mean seed number per square metre per year ± 1 SE from 12 September 1992–12 September 1993.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
A persistent deficiency of vegetation cover in an area of primary succession may result from a lack of seed immigration and/or from a lack of microsites suitable for seed entrapment, germination, and growth. While the existence of safe sites is generally established post hoc through the presence of recently recruited plants (Harper, 1977; Fowler, 1988; Drake, 1992), the widely scattered occurrence of adult Sarcobatus and D. spicata on the playa at Mono Lake demonstrates only that safe sites for plant establishment have existed in the past, providing no insight into their present abundance or distribution. Alternatively, seed immigration is readily quantified through trapping, providing the means to assess this factor as the cause of or contributor to the current lack of recruitment.

Four lines of evidence derived from the large spatial-scale seed rain sampling in this study support the conclusion that seed input does not limit plant colonization on the playa. First, in opposition to our first hypothesis, seeds from the cover-dominant Sarcobatus and Chrysothamnus spp. dispersed to the playa in quantities far exceeding the expected zero to near-zero values. This was especially true for the halophyte Sarcobatus, which dispersed as many as 30.7 ± 4.3 (mean ± 1 SE, N = 32) viable seeds·m·d for 22 d in fall 1993. Second, the proportion of total species trapped on the playa was high: of the 13 seed species obtained throughout the study site, ten of these were found on the playa (Table 3). Third, the only recruitment of Sarcobatus on the playa in the last 40 yr (i.e., since 1957; see Fig. 1) began in 1984–1986 and continued until 1997 when the rising lake inundated the 1984–1986 shoreline berm (K. P. Fort and J. H. Richards, personal observations), yet seed input at this location was the lowest found on the playa (Fig. 3C). Fourth, seeds from an invasive herbaceous species, Bassia hyssopifolia, were recovered from various traps on the playa (Fig. 3E). Plants of this species have become established on the 1984–1986 shoreline berm, yet have not colonized adjacent areas on the playa (K. P. Fort, personal observation).

If seed input is not limiting, why is there a lack of recent recruitment on the playa? The nature of the playa soil plays an important role, especially adjacent to the lake edge (>1600 m in Fig. 1) where the substrate is saturated, anoxic, and encrusted with salt at the surface. In contrast, the higher elevation playa soil has improved drainage and lacks a salt crust (K. P. Fort, personal observation). Nevertheless, soil in this latter area retains high total salts, pH, Na, B, and S, in addition to having low levels of Ca, Mg, P, and N when compared to nonplaya soils (Schaber, 1994; Donovan, Richards, and Schaber, 1997). One or more of these factors strongly suppressed C. nauseosus grown on playa soil in a growth chamber experiment, but Sarcobatus was substantially less affected (Schaber, 1994). The chemistry of the playa soil may therefore limit the colonization of glycophytic species, which grow in and disperse seed from less saline dune sites, but does not seem to entirely explain the slow colonization by the halophyte, Sarcobatus.

Poor seed entrapment may limit Sarcobatus recruitment on the playa. The smooth soil surface, the virtual absence of above ground barriers, and the high windspeed environment (Fig. 2A) promote the secondary dispersal of seeds (i.e., dispersal following the arrival of a seed from the parent plant to a surface; Chambers and MacMahon, 1994) on the upper playa (950–1600 m in Fig. 1). Surface irregularities on the lower playa (>1600 m in Fig. 1) due to salt encrustation probably reduce seed movement. Nevertheless, seeds of Chrysothamnus spp. were collected from traps on the 1984–1986 shoreline berm (data not presented), a location ~1300 m from the nearest seed source (Fig. 3B), illustrating the high dispersal potential of this species in this environment. Although the deep burial of seeds in the playa soil was possible, especially for small seeds (Chambers, MacMahon, and Haefner, 1991), soil seed bank samples collected from the upper playa yielded zero or exceedingly low numbers of seeds (Schaber, 1994).

The bare playa surface resembles the smooth surfaces in other studies, especially with regard to its apparent lack of entrapment microsites. A study of Betula lenta secondarily dispersing across a snow surface revealed relatively high numbers of accumulated seeds in artificial depressions and low numbers on undisturbed snow (Matlack, 1989). In New Mexico, persistently bare floodplains received high seed input, but had low seed retention (Knipe and Springfield, 1972). A wind tunnel study confirmed that seeds moving across relatively smooth surfaces tend to remain stationary for shorter periods of time, providing less time for seeds to imbibe water (Johnson and Fryer, 1992). Biotic factors may also influence the fate of seeds that immigrate to, or are produced on, the playa. Rodent tracks occur under and around the scattered Sarcobatus shrubs, and ants actively collect Sarcobatus seeds (K. P. Fort, personal observations). Ants are known for their active role in the seed dispersal of some species (Levey and Byrne, 1993), especially on infertile soils (discussed in Westoby, Rice, and Howell, 1990). Because ants and rodents are seed predators, frequently consuming a large percentage of available seeds, the occurrence of mutualistic benefits to seeds may be very low or absent (Reichman, 1979; Levey and Byrne, 1993). However, the actual net impact from ants and rodents on the fate of playa seeds is unknown.

The large spatial scale of sampling in this study was used to address our second hypothesis of parallel vegetation and seed rain diversity, but also permitted the detection of dispersal trends that otherwise would have gone unobserved. In the former, vegetation and seed rain diversity did correlate as hypothesized in most cases, but to a weaker degree than we expected. Two factors reduced the gradient of the seed rain diversity relative to the vegetation: first, the predominance of a single species in the seed rain within the diverse vegetation lowered the seed rain diversity at these locations (Fig. 4A, C). Second, seed dispersed to areas of the playa where vegetation was sparse or absent and resulted in some (low) measure of seed diversity and richness where vegetation diversity and richness was near zero (Fig. 4A, B).

An additional large-scale trend included the displacement of the peaks of total and viable seed abundance for Sarcobatus from the peak of percent cover of the parent vegetation (Fig. 3C, D). Strong winds can contribute to such a pattern, as was shown with Cakile edentula growing on a coastal sand dune (Keddy, 1982). In this study, Sarcobatus shrubs growing where dispersal was greatest and cover was low (700–1250 m along the N–S plot axis; Fig. 3C, D) produced dense clusters of large seeds that contrasted sharply to the small seeds and sparse seed production on shrubs where cover was high (300–600 m along the N–S plot axis), farther from the lakeshore (K. P. Fort, personal observation). The lakeward shrubs have higher leaf nitrogen during leaf growth, shoot growth, and flowering than shrubs in the dunes farther from the shore (Donovan, Richards, and Schaber, 1997). Therefore, the observed disparity between the peaks of Sarcobatus vegetation cover and seed dispersal (Fig. 3C, D) probably results from the increased seed production of lakeward shrubs, which have less competition or have greater access to some limiting resource (or both), rather than from wind alone.

The large-scale dispersal pattern of Chrysothamnus spp. is more typical, with the decline of dispersed seed, both total and viable, correlated with the decline in parent vegetation (Fig. 3A, B). The percentage viability of Chrysothamnus spp. also declined with proximity to the lakeshore, a trend that was a likely result of seed sorting based on the increased dispersal ability of lighter, nonviable seeds. The increased dispersal ability of lighter seeds within the same aerodynamic type has been demonstrated in modeling studies (Augspurger and Franson, 1987; Greene and Johnson, 1993), and nonviable seeds have been shown or implicated to disperse farther than viable seeds in other field studies (Stergios, 1976; Chapin et al., 1994).

The predominance of wind-dispersed seed on the playa is comparable to other sites of primary succession, notably the seed dispersal onto a volcanic debris avalanche (Nakashizuka et al., 1993), volcanic pyroclastic flows (Wood and del Moral, 1988; Drake, 1992), and in a region of glacier retreat (Chapin et al., 1994). Seven of the ten different species collected on the playa at Mono Lake had distinct wind-dispersal adaptations (adaptations listed in Table 1, seeds collected on the playa listed in Table 3). Seeds of the remaining three species lacked such adaptations but were very small, albeit too large to be classed as a true wind-dispersal adaptation by virtue of size alone (i.e., <0.05 mg [Fenner, 1985]; seed size and wind dispersal also discussed in van der Pijl, 1982; Frantzen and Bouman, 1989; Willson, Rice, and Westoby, 1990). Nevertheless, these small seeds would be expected to disperse farther during high winds than larger, heavier seeds of similar morphology. In contrast to the types of seeds found on newly formed volcanic soil on Mt. St. Helens, which were either well dispersed from stress-intolerant species or were large and poorly dispersed from stress-tolerant species (del Moral and Wood, 1993), the types of seeds found on the playa at Mono Lake included a high proportion of the wind-dispersed and stress-tolerant species Sarcobatus.

For some plants, the timing of seed release appears to be adapted to temporal variation in dispersal opportunities (discussed in Howe and Smallwood, 1982). In the present study, where wind was the primary agent of dispersal, increased numbers of seeds were not trapped during sampling periods with higher total wind runs. The timing of dispersal does not appear to be adapted to periods of increased wind run for the five dominant species. The generally high daily wind run throughout the year (Fig. 2A) and openness of the vegetation may obviate such an adaptation.

Data collected on widely dispersed seed at Mono Lake yielded both expected and unexpected results. As expected, seed on the playa was predominantly wind-dispersed, a trait common to other sites of primary succession. However, Chrysothamnus spp. dispersed at least 1300 m and Sarcobatus at least 700 m, distances rarely documented in the literature for wind-dispersed species on land (Willson, 1993). Also as expected, the abundance (and sometimes diversity and richness) of the seed rain declined with increased proximity to the lakeshore where vegetation became sparse to nonexistent. However, the abundance of viable seed from the stress-tolerant shrub Sarcobatus was not only surprisingly high on the playa relative to the expected zero to near-zero values, but was also spatially displaced from the parent population. These data, coupled with playa seedbank counts (Schaber, 1994), indicated low seed entrapment, rather than a lack of seed input, as a bottleneck to plant establishment on inhabitable regions of the playa.


   FOOTNOTES
 
1 The authors thank G. Besne, J. Brown, D. Chirman, L. Donovan, T. Fraizer, C. Huang, M. Muller, E. Rowell, J. Schaber, C. Toft, H. West, and Z. Zhu for assistance with field work; M. Hair for use of a hydropneumatic elutriator; L. Donovan, M. Rejmanek, K. Rice, and C. Toft for helpful input on the experimental design and analysis; S. Syrjala for programming help; M. Rejmanek and C. Toft for helpful input on the manuscript; and L. Ford and the Mono Lake Ranger District for cooperation. Funding was provided by USDA NRICGP grants 92-37101-7419 and 94-37101-1144. A USFS Special Use Permit permitted access to the study site. Back

2 Author for correspondence. Back


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Alvarez-Buylla, E. R., and M. Martínez-Ramos.1990Seed bank versus seed rain in the regeneration of a tropical pioneer tree. Oecologia 84: 314–325. [ISI]

Augspurger, C. K., and S. E. Franson.1987Wind dispersal of artificial fruits varying in mass, area, and morphology. Ecology 68: 27–42.

Chambers, J. C., and J. A. MacMahon.1994A day in the life of a seed: movements and fates of seeds and their implications for natural and managed systems. Annual Review of Ecology and Systematics 25: 263–292.

———, ———, and J. H. Haefner.1991Seed entrapment in alpine ecosystems: effects of soil particle size and diaspore morphology. Ecology 72: 1668–1677. [CrossRef][ISI]

Chapin, F. S., III, L. R. Walker, C. L. Fastie, and L. C. Sharman.1994Mechanisms of primary succession following deglaciation at Glacier Bay, Alaska. Ecological Monographs 64: 149–175. [CrossRef]

del Moral, R., and D. M. Wood.1993Early primary succession on the volcano Mount St. Helens. Journal of Vegetation Science 4: 223–234.

Donovan, L. A., J. H. Richards, and E. J. Schaber.1997Nutrient relations of the halophytic shrub, Sarcobatus vermiculatus, along a soil salinity gradient. Plant and Soil 190: 105–117. [CrossRef][ISI]

Drake, D. R.1992Seed dispersal of Metrosideros polymorpha (Myrtaceae): a pioneer tree of Hawaiian lava flows. American Journal of Botany 79: 1224–1228. [CrossRef][ISI]

Fenner, M.1985Seed ecology. Chapman and Hall, London.

Fowler, N. L.1988What is a safe site?: neighbor, litter, germination date, and patch effects. Ecology 69: 947–961. [CrossRef][ISI]

Frantzen, N. M. L. H. F., and F. Bouman.1989Dispersal and growth form patterns of some zonal páramo vegetation types. Acta Botanica Neerlandica 38: 449–465. [ISI]

Greene, D. F., and E. A. Johnson1993Seed mass and dispersal capacity in wind-dispersed diaspores. Oikos 67: 69–74. [CrossRef][ISI]

Gross, K. L., and K. A. Renner.1989A new method for estimating seed numbers in the soil. Weed Science 37: 836–839. [ISI]

Harper, J. L.1977Population biology of plants. Academic Press, New York, NY.

Hickman, J. C., editor.1993The Jepson manual: higher plants of California. University of California Press, Berkeley, CA.

Hill, M. O.1973Diversity and evenness: a unifying notation and its consequences. Ecology 54: 427–432. [CrossRef][ISI]

Howe, H. F., and J. Smallwood.1982Ecology of seed dispersal. Annual Review of Ecology and Systematics 13: 201–228.

Johnson, C. K., and N. E. West.1988Laboratory comparisons of five seed-trap designs for dry, windy environments. Canadian Journal of Botany 66: 346–348. [CrossRef]

Johnson, E. A., and G. I. Fryer.1992Physical characterization of seed microsites—movement on the ground. Journal of Ecology 80: 823–836. [CrossRef]

Keddy, P. A.1982Population ecology on an environmental gradient: Cakile edentula on a sand dune. Oecologia 52: 348–355. [CrossRef][ISI]

Knipe, O. D., and H. W. Springfield.1972Germinable alkali sacaton seed content of soils in the Rio Puerco Basin, West Central New Mexico. Ecology 53: 965–968. [CrossRef][ISI]

Levey, D. J., and M. M. Byrne.1993Complex ant-plant interactions: rain forest ants as secondary dispersers and post-dispersal seed predators. Ecology 74: 1802–1812. [CrossRef][ISI]

Ludwig, J. A., and J. F. Reynolds.1988Statistical ecology: a primer on methods and computing. John Wiley & Sons, New York, NY.

Marlette, G. M., and J. E. Anderson.1986Seed banks and propagule dispersal in crested-wheatgrass stands. Journal of Applied Ecology 23: 161–175. [CrossRef][ISI]

Matlack, G. R.1989Secondary dispersal of seed across snow in Betula lenta, a gap-colonizing tree species. Journal of Ecology 77: 853–869. [CrossRef]

Mono Basin Ecosystem Study Committee.1987The Mono Basin ecosystem: effects of changing lake level. National Research Council. National Academy Press, Washington, D.C.

Nakashizuka, T., S. Iida, W. Suzuki, and T. Tanimoto.1993Seed dispersal and vegetation development on a debris avalanche on the Ontake volcano, Central Japan. Journal of Vegetation Science 4: 537–542. [CrossRef][ISI]

Potvin, C., and M. J. Lechowicz.1990The statistical analysis of ecophysiological response curves obtained from experiments involving repeated measures. Ecology 71: 1389–1400. [CrossRef][ISI]

Reichman, O. J.1979Desert granivore foraging and its impact on seed densities and distributions. Ecology 60: 1085–1092. [CrossRef][ISI]

SAS.1988SAS/STAT user's guide. Release 6.03 edition. SAS Institute, Cary, NC.

Schaber, E. J.1994Salinity and nutrient effects on the germination and growth of Chrysothamnus nauseosus and Sarcobatus vermiculatus in playa dunes around Mono Lake, California. Master's thesis. University of California, Davis, CA.

Silvertown, J. W., and J. L. Doust.1993Introduction to plant population biology. Blackwell Scientific Publications, Boston, MA.

Stallings, G. P., D. C. Thill, C. A. Mallory-Smith, and L. W. Lass.1995Plant movement and seed dispersal of Russian Thistle (Salsola iberica). Weed Science 43: 63–69. [ISI]

Stergios, B. G.1976Achene production, dispersal, seed germination, and seedling establishment of Hieracium aurantiacum in an abandoned field community. Canadian Journal of Botany 54: 1189–1197.

Stine, S.1991Geomorphic, geographic, and hydrographic basis for resolving the Mono Lake controversy. Environmental Geology and Water Sciences 17: 67–83.

Syrjala, S. E.1996A statistical test for a difference between the spatial distributions of two populations. Ecology 77: 75–80.

Toft, C. A.1995A 10-year demographic study of rabbitbrush (Chrysothamnus nauseosus): growth, survival and water limitation. Oecologia 101: 1–12.

van der Pijl, L.1982Principles of dispersal in higher plants. Springer-Verlag, Berlin.

von Ende, C. N.1993Repeated-measures analysis: growth and other time-dependent measures. Pages 113–137 in S. M. Scheiner and J. Gurevitch, editors. Design and analysis of ecological experiments. Chapman and Hall, New York, NY.

Westoby, M., B. Rice, and J. Howell.1990Seed size and plant growth form as factors in dispersal spectra. Ecology 71: 1307–1315. [CrossRef][ISI]

Willson, M. F.1993Dispersal mode, seed shadows, and colonization patterns. Vegetatio 107/108: 261–280.

———, B. L. Rice, and M. Westoby.1990Seed dispersal spectra: a comparison of temperate plant communities. Journal of Vegetation Science 1: 547–562. [CrossRef][ISI]

Wood, D. M., and R. del Moral.1988Colonizing plants on the Pumice Plains, Mount St. Helens, Washington. American Journal of Botany 75: 1228–1237. [CrossRef][ISI]




This article has been cited by other articles:


Home page
 ANN BOT (LOND)Home page
F.-R. LI, T. WANG, A.-S. ZHANG, L.-Y. ZHAO, L.-F. KANG, and W. CHEN
Wind-dispersed Seed Deposition Patterns and Seedling Recruitment of Artemisia halodendron in a Moving Sandy Land
Ann. Bot., July 1, 2005; 96(1): 69 - 80.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Bot.Home page
G. L Dodd and L. A Donovan
Water potential and ionic effects on germination and seedling growth of two cold desert shrubs
Am. J. Botany, August 1, 1999; 86(8): 1146 - 1153.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (24)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Fort, K. P.
Right arrow Articles by Richards, J. H.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Fort, K. P.
Right arrow Articles by Richards, J. H.
Agricola
Right arrow Articles by Fort, K. P.
Right arrow Articles by Richards, J. H.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS