— Madrono, Vol. 46, No. 1, pp. 46-48, 1999 NOTES IMPACT OF A NON-NATIVE PLANT ON SEED ed in areas below the plant before the stemabscised DISPERSAL OF A NATIVE. Salsola tragus L. and where the plant came to rest for a time (i.e., (Mosyakin 1996), or tumbleweed, was introduced obstructions). By contrast, E. deflexum have winged to the United States from Eurasia in the late 1800's seeds carried from the maternal plant by wind cur- (Young 1991). It has subsequently invaded large rents. The dispersal of these seeds depends upon portions ofthe arid western United States and Can- the speed and direction of the wind, which may be ada. The impact that Salsola spp. have upon neigh- altered by plants in the area. Salsola tragus in boring native plant survival during succession has southwestern Wyoming had wind speeds of 13 m/ been the subject ofa number ofstudies (e.g., Lodhi sec one meter above the plant, 11 m/sec 15 cm 1979; Allen and Allen 1988; Johnson 1998). How- above the plant, 9 m/sec 10 cm in front (i.e., up- ever, the impact of the invader upon the seed dis- wind) of the plant, and 2.5 m/sec 10 cm behind persal of natives has not been quantified. Plants (i.e., downwind from) the plant (Allen & Allen such as Salsola spp. that have a shrubby growth 1988). We hypothesized that S. tragus plants pro- form act as barriers that slow windcurrents andtrap vide a barrier to E. deflexum seed dispersal through wind-dispersed seeds (Day and Wright 1989). Sal- slowed wind currents. Eriogonum deflexum plants sola tragus often reaches densities higher than na- may in turn affect S. tragus seed dispersal by fur- tive species on disturbed sites. As aresult, S. tragus nishing a physical obstruction to movement during may impact the seed dispersal of other species to a the tumbling phase. greater extent than natives with a similar architec- On the overburden pile at Castle Mountain Mine, ture. Thus, the presence ofthe non-native S. tragus we selected a flat, 600 m2 plot dominated by S. has the potential to alter the ability of natives to tragus and, to a lesser extent, E. deflexum. We hap- colonize successfully by altering seed dispersion hazardly chose sixty plants representative ofthe en- m patterns and plant survival. tire 600 2 plot: fifteen each oflive S. tragus, dead Castle Mountain Mine near Searchlight, NV, (but attached to the ground by the main stem) S. contains a large (ca. 50 ha) overburden area where tragus, live E. deflexum, and dead E. deflexum. unusable excavated material from deep within the Most plants chosen were at least 50 cm from their mountain is piled. The surface consists of heavily nearest neighbor (dead E. deflexum were occasion- compacted rock fragments with nutrient-poor (total ally as close as 20 cm to their nearest neighbor). In Kjeldahl nitrogen <1.5 mg/g), alkaline (pH = 8.0) September 1996 we collected seed and plant litter soils that have little organic matter (<1.59%) and within a 10 cm X 15 cm quadrat under each plant clay (<1.0%) (Walker unpublished data). The most and from the nearest plant-free clearing within 25 common colonizers include the non-native annual cm to 100 cm east of each focal plant's eastern S. tragus and the more diffusely constructed native canopy edge. We used the difference between each annual Eriogonum deflexum Torrey (Hickman canopy sample and its adjacent open area for anal- 1993), or flat-topped buckwheat. Although both ysis. This approach adjusted for differences in spe- species have wind-dispersed seeds, the two plants cies substrate preferences on a microsite scale (e.g., differ in their specific mechanisms of seed dispers- living E. deflexum were more common in seed- al. Salsola tragus disperses its seeds after the main catching rocky areas while living S. tragus were stem abscises at the base. The wind subsequently more common on sandy substrates). Litter was sep- pushes the elliptical plants,jarring seeds loose from arated from the inorganic soil by sieving each sam- leaf axils as the plants bounce on the ground over ple and floating the organic matter. Each litter sam- a considerable distance. Stallings et al. (1995) ple was dried and weighed. Due to the time-con- found that Salsola tragus seeds were evenly dis- suming nature of seed counting, five random sam- tributed across the landscape if they were dropped ples from each treatment were selected for seed during the tumbling phase. Seeds were concentrat- counts. All seeds were identified for each sample, Tab—le 1. Descriptive Statistics for All Status and Species Combinations Adjusted for Microsite (UnderCan- opy Adjacent Open Area) [Mean (SE)]. Seed densities expressed as numberofseeds percm2. Littermassexpressed as grams dry weight per cm2 . Dead E. deflexum Live E. deflexum Dead S. tragus Live S. tragus Seed densities E. deflexum 1.35 (1.20) 4.30 (1.93) 9.62 (4.31) 2.32 (1.24) S. tragus 0.23 (0.19) 0.09 (0.09) 4.05 (2.78) 3.85 (1.30) Litter mass 0.0010 (0.0005) 0.0139 (0.0024) 0.6720 (0.2390) 0.0315 (0.0053) 1999] NOTES 47 Table 2. Summary from Two-Way ANOVA with Sta- tus ofPlant (Dead or Alive) and Species (E. deflexum or S. tragus). Seed—densities and litter adjusted for mi- crosite (underplant adjoining open area) and natural log 14 transformed prior to analysis. F ratios are reported, with 12 J—-- LDievaindgPPllaannttss a*s*te=risPks<de0.n0o1t0i)n.g significance level (* = P < 0.050; 10 deflexum S. tragus seed seed *S Source of variation df density density df Litter E-a CO V<)D SStpaetcuises 00..0010 100..1747** 1 5106..9159**** 1 Status X Species 0.40 2.61 18.81** 1 Error 56 Total 59 Eriogonumdeflexum Salsolatragus PlantSpecies but seeds other than S. tragus and E. deflexum were very rare and not included in the analysis. The data were analyzed with a two-way ANOVA (Minitab 1991) which included main effects of species and status (dead or alive) and the interaction between them. Dead Plants 6 - Living Plants Results Samples from open areas had less litter (mean ± 5 - SE; 0.003 ± 0.001 vs. 0.032 ± 0.007 g. dry wt./ S1o c±m12).,3f9esweeerdEs./cdme2f)l,eaxnudmfseeweedsr(S.1.t3r0ag±us0.s5e4edvss.(50..6199 O 0) ± 0.08 vs. 2.25 ± 0.84 seeds/cm2) than areas under </} wd> plant canopies, as expected. After adjusting the un- 75 COJfc. der plant sample for microtopography using the open sample, the presence of S. tragus plants re- sulted in E. deflexum seed densities statistically equal to those under living E. deflexum plants, re- gardless of plant status (Tables 1 and 2; Fig. 1A). Eriogonumdeflexum Salsolatragus These results suggest that S. tragus plants served PlantSpecies as an effective trap forE. deflexum seeds. However, S. tragus seed density was more than 16 fold great- er under S. tragus than under E. deflexum (Tables 1 and 2; Fig. IB), indicating that the native E. de- C. flexum does not provide a major barrier to tumbling 0.10 S. tragus. Plant status, species, and the interaction 0.09 -•— Dead Plants between them were all significant forlitterdata (Ta- 0.08 Living Plants bles 1 and 2; Fig. 1C). Thus S. tragus plants were also more effective than E. deflexum in catching 0.07 and/or retaining litter under the canopy. wu) Eo 0.06 The data suggest that E. deflexum seeds under E. 0.05 deflexum fall from the plant and are blown away, i b 0.04 H while E. deflexum seeds under S. tragus have been _JTJ 0.03 captured by the slowed wind currents (Fig. 1A) and 0.02 0.01 0.00 Fig. 1. Interactions between plant status and species for -0.01 Eriogonumdeflexum Salsolatragus a)Eriogonumdeflexum seeds, b)Salsola tragusseeds,and PlantSpecies c) litter [mean ± SE]. Means adjusted for microsite by subtracting adjacent open sample from canopy sample. Numbers shown in Table 1. MADRONO 48 [Vol. 46 perhaps held by litter under dead S. tragus (Fig. pete with adults forwaterandnutrients. Salsolakali 1C). Figure IB indicates S. tragus seeds under S. has an extensive root system and efficient uptake tragus plants probably came from the plant above of phosphorus (Itoh and Barber 1983). In addition, them, since E. deflexum had very few S. tragus S. kali can significantly alter soil nutrient concen- seeds under their canopies (Table 1). We conclude trations in mixed culture, perhaps due to its rapid that S. tragus has a significant effect upon the dis- growth rate (Allen 1982). Salsola tragus can even persion ofE. deflexum seeds, butE. deflexum plants make phosphorus more available to neighboring do not affect the dispersion of S. tragus seeds. plants through a high oxalate concentration in can- However, the time period for which this pattern is opy leachates (Cannon et al. 1995; Hageman et al. obtainable may be strongly influenced by tumble- 1988). Clearly, Salsola spp. affect neighboring na- weed abscission laterin the season, apossibilitynot tives, but the net effect of this interference is un- addressed in this data set. known. This study has shown that Salsola spp. can Further work is needed to identify the overall concentrate native plant seeds under the Salsola impact that the seed-catching function of the alien spp. canopies, where germination and growth may S. tragus has upon native seedling establishment then be either facilitated or inhibited. Further un- and plant success. We showed that a native plant's derstanding of the effects ofSalsola spp. on native seed distribution can be affected by Salsola tragus. colonizers will enhance efforts toreintroducenative Other work has indicated that Salsola spp. can pos- species to damaged ecosystems. itively or negatively interact with nearby natives — throughout the plants' lives. Allen andAllen (1988) Cheryl H. Vanier and Lawrence R. Walker, De- proposeSalsola kalimay facilitatenativegrass seed partment of Biological Sciences, University of Nevada, colonization and establishment but may later com- Las Vegas, NV 89154.