Seed dispersal, seed entrapment, and seedling recruitment in a temporally variable desert playa

Natural establishment of seedlings in desert playas with temporally variable precipitation hinges on many factors, including seed production, seed dispersal, seed entrapment, seed germination, and seedling survival. We investigated natural seed dispersal patterns and the effects of surface texture, wind barriers, resource availability (water and nutrients), and protection from herbivory on seedling establishment in a desert playa. We hypothesized that the seed rain would be consistent throughout the year and that seedling establishment would improve in resource-amended plots with barriers to wind dispersal. Contrary to our hypotheses, seed flux peaked seasonally during the winter, and fertilization had no consistent effect on seedling establishment. Seed availability for dispersal correlated with precipitation in the previous water year, whereas seedling recruitment was greatest when current-year precipitation during the spring germination and summer seedling growth periods was high. Gravel and barrier surface treatments contained more surviving seedlings than other surface treatments. Comparison with plots located outside of herbivore exclosures, however, showed that greater seedling presence in gravel plots may be due somewhat to protection from herbivory provided by the gravel, rather than simply to the greater seed-trapping quality of the gravel. With abundant seed availability, application of surface treatments, like coarse gravel, combined with increased seasonal water availability could lead to improved shrub establishment from seed in desert playas.     

Lagomorphs and the dispersal of seeds into communities dominated by exotic annual weeds

Large areas of western rangeland are presently dominated by alien annual weeds such as Bromus tectorum (cheatgrass). These communities resist succession to perennial communities primarily because the annuals are competitively superior to establishing perennial seedlings and they promote fires that favor weeds over perennials. Succession may be further slowed, however, by low rates of seed dispersal into annual grasslands. We investigated the role of lagomorphs ( Sylvilagus nuttalli , Nuttall's cottontail; S. audubonii , desert cottontail; and Lepus californicus , black-tailed jackrabbit) in seed dispersal across an ecotone: 50 m into woodland, border, and 20 m, 50 m, and 100 m into grassland. We searched pellets for juniper seeds visually and for any other species through germination from crushed pellets after cold, moist stratification. Pellets were not even distributed across transects, but there was no trend with respect to position of transect. Juniperus osteosperma (Utah juniper) was the most abundant seed. Both the number of juniper seeds and the proportion of pellets with juniper seeds decreased steadily from a high in woodland to absence at 100 m into grassland. Only 2 dicot seedlings emerged from pellets, 1 Salsola pestifer and 1 unknown that died prior to identification. Consequently, there was little seed movement into grassland; 72% of all seeds were collected from either woodland or border transects. Lagomorphs apparently do not effectively replenish the native perennial seed pool of cheatgrass-dominated disturbances at Dugway.     

Moistened seeds increase rodent trap success

Seed moisture has been shown to influence the rates of seed cache removal by rodents. Although the precise mechanism is not known, this knowledge might prove useful in field applications. We examined whether moistened bait would increase trap success in desert rodent populations. We placed traps 15 m apart in grids within a 500-ha study area and randomly baited traps with either dry or moistened seeds. We found that traps baited with moistened seeds had 34.9% higher success than traps baited with dry seeds ( n = 190, Χ 2 = 5.389, df = 1, P = 0.020). Our results suggest that application of water to dry seed bait can lead to increased trap success for desert rodents.     

Seed dispersal in hybrid Salsola

The study population is a group of hybrids of Salsola iberica × S. paulsenii . The plants show great variation in phenotypic characteristics, and the character complexes of the parent species seem to have largely disappeared. Seeds are dispersed both locally, from intact plants, and by tumbling. Tumbling is the dominant form of dispersal in most individuals studied. Most locally dispersed seeds fall between .5 and 1.5 m from the parent plant. Despite the dominance of tumbling, large numbers of seeds are dispersed locally and probably serve to maintain the population while those dispersed by tumbling introduce the plants into new areas.     

Seed production and seedling establishment of a Southwest riparian tree, Arizona walnut ( Juglans major )

A four-year study of five populations has revealed influences on seed production and seedling establishment of the Southwest riparian tree Juglans major . Germination is abundant after production of large seed crops (masts), but masts are produced infrequently. Within years, germination is stimulated by summer rains, enabling seedlings to establish on riparian terraces as well as streambanks. Traits such as capacity for dormancy during summer drought allow some seedlings to survive on terraces, but abundant rainfall is essential for high rates of seedling success. Ranges of moisture tolerance vary among seeds collected from different populations, suggesting that ecotypes may exist between riparian sites with dissimilar moisture regimes. Population-based differences are associated, in part, with differences in seed size.     

Forest-floor treatments in Arizona ponderosa pine restoration ecosystems: no short-term effects on plant communities

Leaf litter accumulation during fire exclusion and increases in tree density in postsettlement southwestern Pinus ponderosa forests may limit the establishment of understory vegetation. We performed an experiment in P. ponderosa forests of northern Arizona to ascertain plant community responses to forest-floor scarification and Oi removal on thirty-six 100-m 2 plots overlaid on an existing ecological restoration experiment that involved tree thinning and prescribed burning. Constrasting with findings from many other forest types, forest-floor treatments had no effect on community diversity or composition during the 2-year experiment. S?rensen similarities were as high as 97% between posttreatment years within treatments; and successional vectors also provided little indication that treatments may appreciably affect longer-term successional trajectories. Lack of response to these fairly drastic treatments is surprising given these forests' exceptionally heavy Oi horizons and large proportions of conifer litter. Apparently shading, belowground competition for water or nutrients, or other tree-associated factors more strongly limit understory communities than does leaf litter. Based on sparse A-horizon seed banks averaging <300 seeds ? m –2 and limited aboveground vegetation, we hypothesize that seed shortages, particularly for native perennials, also partly precluded a treatment response. Because extensive unvegetated areas at these restoration sites may be colonized by exotics, conservative management strategies could include testing the seeding or outplanting of desirable native species as an option for filling unoccupied microsites. Reporting of "no treatment effect" experiments such as this one is important to avoid biasing meta-analyses, as is future research to clarify combinations of factors limiting understory communities. Increased understanding of these limiting factors may lead to identification of other treatments that promote recovery of native species during ecological restoration in this region.     

Comparing the effects of granivorous rodents on persistence of Indian ricegrass ( Oryzopsis hymenoides ) seeds in mixed and monospecific seed patches

In desert environments seeds are often heterogeneously distributed in small patches that vary in number of seed species and in seed density. Because seed harvest by rodents is often density dependent (a larger proportion of seeds is removed from high-density seed patches than from low-density patches), the proportion of residual or post-harvest seeds should be greater in low-density patches. In addition, seed preference can affect harvest. We tested whether the residual proportion of a highly preferred seed (Indian ricegrass, Oryzopsis hymenoides ) was less when in a seed patch with a 2nd species (mixed-species patch) than when in a monospecific seed patch. We predicted that the increased overall seed density due to the presence of 2 species in a patch would result in a lower residual proportion of ricegrass seeds in the mixed-species seed patches than in the monospecific patches. As predicted, the residual proportions of Indian ricegrass seeds were less each time ricegrass was paired with one of 6 other species in mixed-species patches. Similarly, the residual proportion of each of those 6 species was less when paired with ricegrass than when in a monospecific patch. We speculate on the potential implications of these results for the population dynamics of plant species and the physical structure of plant communities.     

Influence of yellow starthistle ( Centaurea solstitialis ) on small mammals in central California

Yellow starthistle ( Centaurea solstitialis ) is native to Europe and Asia and has quickly invaded disturbed grasslands and rangelands in the western United States. The purpose of our study was to determine the effects of starthistle on rodents at Beale Air Force Base, Yuba County, California, by examining rodent species diversity, species abundance, age structure, and reproductive condition among locations with low, medium, and high percent cover and height of starthistle. Listed in decreasing order of abundance, the rodents were house mouse ( Mus musculus ), California vole ( Microtus californicus ), deer mouse ( Peromyscus maniculatus ), western harvest mouse ( Reithrodontomys megalotis ), and roof rat ( Rattus rattus ). Indices of diversity did not differ among starthistle cover categories. Regression analyses showed that Reithrodontomys megalotis was more abundant in high-cover starthistle plots, with 90.5% of captures occurring in at least 40% starthistle cover. Significant differences were found in capture rates, reproduction, and age between season only. To manage for all rodent species, an area with medium percent cover and height of starthistle will likely provide adequate protective cover.     

Establishment, growth, and early survival of woody riparian apecies at a Colorado gravel pit

Presence of a wetted edge during the period of seedfall was an effective predictor of suitable establishment (defined as germination and survival to the 1st autumn) locations for Populus deltoides subsp. monilifera, Salix amygdaloides, S. exigua, and Tamarix ramosissima seedlings during 3 successive years of a gravel pit revegetation project in Fort Collins, Colorado. At locations predicted to be suitable for establishment, position within the pit (possibly reflecting additional moisture provided by seepage) was a significant factor in determining whether establishment actually occurred. Cover of herbaceous species, which became established at the same time as, or after, woody seedlings, was positively related to probability of establishment for 8 of 11 species-year combinations, probably reflecting more favorable moisture conditions at certain locations. Herbaceous cover also was positively related to seedling height at the end of the 1st summer of growth for 9 of 11 species-year combinations. Neither establishment nor 1st-summer growth was consistently related to overall decline in the water table as estimated by the drop in surface-water level during the growing season. Flooding in the 1st spring after establishment was negatively related to subsequent survival for 5 of 8 species-year combinations. The 4 species established at different elevations in the pit, depending on location of the wetted edge during the period of seedfall, and there was no evidence that differential mortality subsequently altered their distribution along the elevation gradient. However, the primary objective in this study was to restore native woody species, and we attempted to maintain conditions conducive to meeting this objective. Differential postestablishment mortality may be more important in structuring the riparian community in more rigorous riverine environments.     

Weed control and soil amendment effects on restoration plantings in an Oregon grassland

The restoration of perennial grasslands in western North America often depends on effective weed control. We took advantage of a grassland restoration site on the Nature Conservancy's Agate Desert Preserve in southern Oregon (TNC 1997), where 3 sites had been previously burned, mowed, or both. At these sites we carried out a series of controlled, replicated experiments designed to test the effectiveness of 3 weed control measures: (1) sawdust, (2) glyphosate herbicide, and (3) herbicide plus an alfalfa mulch. All plots were seeded with a mix of 3 native perennial grasses. The soils of the 3 areas differing in previous vegetation management were similar, with the exception of total available soil nitrogen, which was significantly lower in the 2 burned sites. The sawdust treatment reduced total available soil nitrogen, but only in the unburned site, and only in the first few months after application. In all 3 areas the alfalfa mulch significantly increased total available soil nitrogen. However, none of these soil nitrogen differences significantly affected the success of weeds or planted perennial grasses. The herbicide treatment reduced exotic annual grasses and forbs and greatly increased the success of native forbs and the planted perennial grasses. The herbicide increased both initial establishment of the native grasses and their absolute cover and biomass. These results suggest that neither nitrogen impoverishment nor nitrogen enrichment was a useful restoration technique at this site, but weed control by herbicides can be of considerable assistance in restoring native perennial grasses.     

Influence of experimental habitat manipulations on a desert rodent population in southern Utah

This paper addresses how habitat manipulations in a black sagebrush ( Artemisia nova ) -dominated area, John's Valley of southern Utah, affected resident desert rodent populations. Rodents studied included the deer mouse ( Peromyscus maniculatus ), Great Basin pocket mouse ( Perognathus parvus ), sagebrush vole ( Lagurus curtatus ), Ord's kangaroo rat ( Dipodomys ordii ), and least chipmunk ( Eutamias minimus ). The experimental design involved analyses of treatment and control (nontreatment) plots rather than pre- and posttreatment of all plots. Habitat manipulations emphasized cutting of shrubs (rotobeating), treatment of plants with a herbicide (2,4-D), and reseeding with a mixture of grasses, forbs, and shrubs. Posttreatment trapping indicated the deer mouse was the most abundant rodent in treatment and control plots. Data indicate the prescribed habitat treatments had no significant negative affects on the deer mouse demes on the control or treatment plots. Habitat treatments may have negatively impacted recruitment in pocket mice. Least chipmunks were not captured in plots treated by rotobeating. Our habitat manipulations may have contributed to interspecific competition in this rodent community through the reduction of both food and cover.     

Tamarix aphylla : a newly invasive tree in southern Nevada

In the southwestern United States, the nonnative athel pine ( Tamarix aphylla ) was presumed to be sterile and therefore not as likely to spread as its widely distributed, nonnative congener, T. ramosissima . However, at Lake Mead National Recreation Area (LMNRA) in southern Nevada, populations of T. aphylla have recently spread beyond their limited pre-1990 distribution and now form extensive monospecific stands. Over a 3-year period, we quantified seed production and germination from 60 T. aphylla trees at LMNRA. The annual mean seed production period was 50.6 days, and the mean potential germination (under laboratory conditions) was 22%, indicating that T. aphylla trees at LMNRA are capable of sexual reproduction in southern Nevada. No seeds germinated in field experiments, apparently because of high soil salt levels. However, seedling regeneration is becoming increasingly common at LMNRA. Tamarix aphylla trees occupied a distinct zone along the shoreline of Lake Mead, above T. ramosissima in elevation and below native Larrea tridentata communities, suggesting either competitive exclusion or differential resource utilization. The T. aphylla zone tended to have the highest mean values for total vegetation cover, leaf litter depth, soil salinity, soil moisture, pH, total Kjeldahl nitrogen, and soil organic matter. The capacity for sexual reproduction of this alien plant, combined with a suite of characteristics shared with the invasive T. ramosissima (e.g., drought tolerance and copious saline leaf litter), makes T. aphylla a potentially invasive species along the shores of LMNRA and other mesic areas in the desert Southwest.     

Effect of soil microtopography on seed bank distribution in the shrub-steppe

The relationship between soil surface cryptogamic crusts and seed banks was investigated in the shrubsteppe in the Lower Columbia Basin. Seventy-four percent of the seeds in a disturbed bunchgrass community were found in crevices bordering cryptogamic crust polygons. In a sagebrush/bunchgrass community, 89% of the seeds were found in crevices. In a disturbed bunchgrass community, Bromus tectorum seeds were found in both the seed bank and aboveground vegetation communities. Bromus tectorum seeds were located in the seed bank of a sagebrush/bunchgrass community, although it had a minor presence in the aboveground community. Seeds of Artemisia tridentata Nutt. were not found in either the bunchgrass or sagebrush/bunchgrass communities. The high number of seeds found in crevices bordering the cryptogamic crust suggests that crevices play a role in determining the small-scale distributional pattern of shrub-steppe plants at the Fitzner-Eberhardt Arid Lands Ecology Reserve.     

Native plant community composition and genetic diversity associated with long-term weed invasions

Many studies have assessed genetic changes in exotic plant species in their native and introduced ranges, but none have focused on genetic variation in native plant species in response to exotic invasion. We examine characteristics of native plant communities within and outside old (> 25 year) invasions of Acroptilon repens (Russian knapweed) and Cardaria draba (hoary cress). We also document genetic variability of 4 native grass populations ( Hesperostipa comata [needle and thread], Achnatherum hymenoides [Indian ricegrass], Sporobolus airoides [alkali sacaton], and Poa secunda [Sandberg bluegrass]) from 2 areas: adjacent to and within weed invasions. Native plant species richness and diversity did not differ between invaded and noninvaded areas. Inter-simple sequence repeat (ISSR) analysis of individual native perennial grasses of each of the 4 species suggests that populations exposed to long-term coexistence with exotics may differ from adjacent noninvaded populations. We suggest that future research efforts should focus on intraspecific diversity of native plant species to identify possible candidates for restoration following weed control.     

Vegetation response to fire and postburn seeding treatments in juniper woodlands of the Grand Staircase-Escalante National Monument, Utah

We compared 3 naturally ignited burns with unburned sites in the Grand Staircase-Escalante National Monument. Each burn site was restored with native and nonnative seed mixes, restored with native seeds only, or regenerated naturally. In general, burned sites had significantly lower native species richness (1.8 vs. 2.9 species), native species cover (11% vs. 22.5%), and soil crust cover (4.1% vs. 15%) than unburned sites. Most burned plots, seeded or not, had significantly higher average nonnative species richness and cover and lower average native species richness and cover than unburned sites. Regression tree analyses suggest site variation was equally important to rehabilitation results as seeding treatments. Low native species richness and cover, high soil C, and low cover of biological soil crusts may facilitate increased nonnative species richness and cover. Our study also found that unburned sites in the region had equally high cover of nonnative species compared with the rest of the Monument. Cheatgrass ( Bromus tectorum ) dominated both burned and unburned sites. Despite the invasion of cheatgrass, unburned sites still maintain higher native species richness; however, the high cover of cheatgrass may increase fire frequency, further reduce native species richness and cover, and ultimately change vegetation composition in juniper woodlands.     

Demographic and environmental relations of two rare Astragalus species endemic to Washington County, Utah: Astragalus holmgreniorum and A. ampullarioides

Two endemic locoweeds of Washington County, Utah, Astragalus holmgreniorum and A. ampullarioides , were recently listed as federally endangered plant species (Federal Register 2001). They both occur in few, small populations surrounded by increasing human activity in the St. George, Utah, area. Demographic and habitat studies have been ongoing since 1992 and results of those studies are presented here. Astragalus holmgreniorum (Holmgren locoweed) is presently limited to 1 larger population and 2 smaller, isolated populations on the eastern and western edges of its range. It is a short-lived perennial that grows primarily on the Virgin limestone member of the Moenkopi Formation. Density of A. holmgreniorum over the 10-year study averaged about 2 plants per m 2 . The number of living plants of this species perhaps never exceeds 10,000. In drought years A. holmgreniorum populations are as much as 95% smaller than in years with adequate water, and few plants produce flowers that successfully contribute to the seed bank. Astragalus ampullarioides (Shivwits locoweed) occurs in 4 populations where plant density fluctuated between 0.5 and 4.8 per m 2 during our study period (average = 2.6). This locoweed is restricted to clay soils of the Chinle Formation. Direct threats to both Astragalus species include widespread urbanization, road construction, population fragmentation, and off-highway vehicle use. Indirect threats include competition from aggressive, introduced annual species such as Bromus rubens, B. tectorum, Malcolmia africana , and Erodium cicutarium . Our data show that 3 of 4 species most closely associated with these Astragalus species are introduced. Studies of seed banks and reproduction biology, as well as continued monitoring of these Astragalus species, are critically needed. Studies of the consequences of competition from alien and native species on seedling establishment of these locoweeds are especially needed.     

Seedling size and survival for Chrysothamnus nauseosus

Seedling size and survival in relation to summer drought were examined for Chrysothamnus nauseosus growing under field and greenhouse conditions. In the field, summer survival rates were less than 2% annually for the three years monitored. The effect of initial seedling height on subsequent survival was examined in both the field and greenhouse by grouping seedlings into live and dead categories on each census date and comparing initial heights for seedlings in these categories. For a majority of the census dates, the initial height of surviving seedlings was greater than the initial height of those that subsequently died (significant differences ranged from 1 to 8 mm), indicating that seedlings that were taller at the initiation of the drought period had a higher probability of survival. In the greenhouse, taller seedlings had greater shoot and root biomass and rooting depth. Seedlings that are larger (i.e., taller and have greater aboveground biomass) in late spring appear to have a higher probability of surviving summer drought due to greater rooting depth and hence increased access to moisture in deeper soil layers. Seed availability and safe sites for germination were probably not limiting since large numbers of seedlings successfully germinated in a patchy pattern during the study period. Seedling size and probability of survival were not related to either seedling density or the distance to nearest seedling neighbor. Survival through summer drought appears to be the main limitation to seedling recruitment in this population.     

Historic expansion of Juniperus occidentalis (western juniper) in southeastern Oregon

The chronology of Juniperus occidentalis (western juniper) expansion in eastern Oregon, the effect of plant canopy and interspace on J. occidentalis seedling establishment and growth rates, and the age of J. occidentalis maximum reproductive potential were determined. Measurements were recorded in twenty-two 0.4-ha plots established in sagebrush-grassland communities and six 0.1-ha plots in Populus tremuloides (quaking aspen) communities. J. occidentalis began increasing during the 1880s in stands containing > 130 yr old. Relatively steady establishment ensued into the 1950s and then began to progress at a geometric rate in the 19602. J. occidentalis encroachment into aspen stands began between 1910 and 1920. The largest proportion of juvenile trees established beneath Artemisia species in sagebrush-grassland communities. J. occidentalis trees appeared to reach full reproductive potential at > 50 yr of age. The ratio of male:female trees increased from 1.7 in scattered J. occidentalis stands to 3.8 in closed stands. The initiation of J. occidentalis encroachment during the late 1800s coincides with optimal climatic conditions for Juniperus berry production and establishment, reduced fire-return intervals, and heavy livestock grazing. The accelerated increase in J. occidentalis expansion since 1960 may be due to the continued absence of fire, abundant woody plant cover, and the large increase in J. occidentalis seed production.     

Cyanobacteria and cyanolichens: Can they enhance availability of essential minerals for higher plants?

In both field and greenhouse studies, cyanobacterial and cyanolichens of cold-temperature deserts often enhance growth and essential uptake by associated herbs. That effect is associated with better seedling establishment and larger seedlings. The following are possible mechanisms for these effects: (1) the microbiota concentrate essential elements in available forms in soil surface layers, (2) the microbial surface covers are usually darker colored than the soil itself and produce warmer soils during cool seasons when soil water is most available, (3) the gelatinous sheaths of several cyanobacterial genera common on alkaline deserts contain chelating compounds, and (4) conditions that favor persistent microbial growths on soil surfaces also favor maintenance of larger populations of microorganisms that form mycorrhizal and/or rhizosheath associations with seed plants. There is evidence that associated animals may be nutritionally benefited by the enhanced mineral content of forage plants growing in well-developed cyanobacterial crusts.     

Reproductive biology, hybridization and flower visitors of rare Sclerocactus taxa in Utah's Uintah Basin

We studied the mating system and flower visitors of 2 threatened species of Sclerocactus (Cactaceae) in the Uintah Basin of eastern Utah—an area undergoing rapid energy development. We found that both S. wetlandicus and S. brevispinus are predominantly outcrossed and are essentially self-incompatible. A third presumptive taxon (undescribed; here called S. wetlandicus -var1) is fully self-compatible but cannot produce seeds unless the flowers are visited by pollinators. We found tentative evidence for pollinator limitation of fruit set in S. brevispinus but not in the other taxa. Pollinators are largely species of native ground-nesting bees in the subfamily Halictinae. These bees are indiscriminate collectors of pollen and nectar from Sclerocactus flowers and other plant species. The well-being of Sclerocactus populations requires that land managers ensure the preservation of halictid bee habitat, which includes both nesting sites and a diversity of floral resources. These bees are also undoubtedly the pollen vectors responsible for the production of hybrids in areas where S. wetlandicus and S. brevispinus meet. We experimentally confirmed that reciprocal crosses between S. wetlandicus and S. brevispinus are indistinguishable from each other or from pure crosses in fruits/flower, seeds/fruit and seed weights. We suggest that hybridization preceded energy development, that it attests to the ongoing process of evolution in the Sclerocactus group, and that it should not be suppressed by anthropogenic activity.