MADRONO, Vol. 56, No. 4, pp. 229-237, 2009

SOIL AND COMMUNITY CHARACTERISTICS ASSOCIATED WITH ()

GEORGE L. VOURLITIS, JULIE MILLER, AND KARl COLER Department of Biological Sciences, State University, San Marcos, CA 92096 [email protected]

ABSTRACT Hazardia orcuttii (A. Gray) Greene is a 5-10 dm tall perennial shrub that is native to coastal sage scrub communities of southern California and northern . This species was listed as threatened by the California Department of Fish and Game in 2002 and is a federal candidate species, and the only known population in the U.S. is on a 1.6 ha mesa located in Encinitas, California. Very little is known about the general ecology of this species, thus, the goal of this .research was to characterize the basic soil physical and chemical properties and community characteristics associated with this species. Research was conducted between January 2004 and July 2005 in 12.56 m 2 randomly-located plots tl:mt either contained or lacked H. orcuttii. Soil in plots containing H.. orcuttii had significantly higher clay, soil 0rganic matter, total N, and soil moisture content than plots lacking H. orcuttii, while plots lacking If. ,orcuttii had significantly more surface litter content. Significant differences were also observed in plant species abundance between plots containing and'1acking H. orcuttii, indicating fundamental differences in plant community composition associated with patches . of H. orcuttii. Our data support the notion that H. orcuttii is a soil endemic; however, it is unclear whether H. orcuttii prefers soil richer in clay or is restricted to these soils because of other factors. Given the restricted nature of H. orcuttii, and the proximity of the extant population to residential areas, habitat protection from human degradation and fire should be a high priority.

RESUMEN Hazardia orcuttii (A. Gray) Greene es un arbusto perenne que mide 5-10 dm de alto que es nativo a Ia comunidad perteneciente de Ia salvia chaparro costeno en el sur de California y ella parte fronterizo de Baja California Norte. Esta especie fue enumerada como una que eata bajo de amenaza por el Departamento de California pescaderia y casa en el 2002, y Ia unica poblacion en los Estados Unidos, esta en una mesa de 1.6 ha ubicado en Encinitas, Calif. Poco se sabe sobre Ia ecologia general de esta especie, asi, el fin de esta inverrstigacion era de caracterisar las propiedades basicos fisicos y quimipos de Ia tierra y las carecteristicas planta associadas con esta comunidad de esta especie. Invistigaciones fueron condujidos entre Enero 2004 y Julio 2005 en 12.65 m 2 en parcelas establecidas al azar unos conteniendo y otros careciendo H, orcuttii. Tierra en parcelas conteniendo H. orcuttii tenian significativamente alto niveles de barro, mater organica del suelo, nitrogeno total, y contenido de humedad del suelo que las parcelas careciendo H. orcuttii tenian significatimante mas contenido de revoltura al superficie. Diferencias significativas fueron observados en Ia abundancia de especie de plantas entre parcelas conteniendo y careciendo de H. orcuttii, indicando diferencias esenciales en el compuesto de la comunidad de las plantas asociado con parches de H. orcuttii. Nuestros datos apoyan la nocion que H. orcuttii es endemica de Ia tierra sin embargo, no es claro si H. orcuttii prefiere tierra rico en barro o esta limitado a esta tierra por otros elementos. Da~o por Ia naturaleza limitado de H. orcuttii, y la cercania de la proximidad de una poblaciones que existe en areas residenciales., proteccion de los habitos degradantes causados por Ia humanidad y el fuego derian ser de alta prioridad. Key Words: Asteraceae, biodiversity, chaparral, coastal sage scrub, human impacts, soil, threatened species.

Habitat loss reduced the extent of coastal sage Specimens collected between 1920 and 1985 scrub by 72% from 1970-1990 (Pryde 1992), indicate that the distribution ofH. orcuttii ranged causing some plant and animal species to become from Encinitas, California to Punta. Colonet, threatened or endangered. One such species, Baja California, . The current distribution Hazardia orcuttii (A. Gray) Greene (Orcutt's is uncertain (Gogol-Prokurat and Osborne 2002) Hazardia), is a 5-10 dm tall resinous evergreen and only two of the 13 previ·ous1y documented shrub in the Asteraceae family that is native to Mexican populations have been located as of maritime sage scrub-chaparral communities of 2004. This plant naturally occurs in only one southern California and northern Baja California documented location in the , on a (Hickman 1993) and was listed by the California 1.6lia m~:;;a (elev. 90-120 m) approximately 5 km Department of Fish and Game as threatened in from tJi~ Pacific coast near Lux Canyon in the August 2002 (Gogol-Prokurat and Osborne 2002). Man9h.~ster Conservation Area in Encinitas, 230 [Vol. 56

California. Moreover, plants are distribiftted in wettest month is February with 64 mm ofrainfall patches that occupy an approximately '&'>.15 ha and the driest month is August with 0.4 mm area located in the SW comer of the mesa. The rainfall. number of plants in this population has been estimated to be 50-700 (Oberbauer 1981; Gogol­ Field Sampling and Data Collection Prokurat and Osborne 2002; Vourlitis et al. 20Q6). The population appears to be long Field plots consisting of 12.56 m 2 permanent established (based on field observations of plant circular quadrats were randomly established in size and woodiness), and voucher specimens from sub-sites containing H orcuttii (n = 13 plots; 1979 indicate that the population has been hereafter referred to as "H. orcuttii plots") and in established at Lux Canyon for at least 30 yr. adjacent sub-sites lacking H orcuttii (n = 10 Hazardia orcuttii occurs in a sage scrub-chaparral plots; hereafter refer-red to as "non-H. orcuttii habitat along with other perennial species such as plots"). As mentio:tfed above, sub-populations of Rhus integrifolia (Nutt.) W. H. Brewer & S. H. orcuttii are restricted to a 0.15 ha portion of Watson., Adenostoma fasciculatum Hook. & the mesa top. Within this area circular plots were Am., and Artemisia californica Less. (nomencla­ established within patches containing H orcuttii ture according to Hickman 1993). and patches lacking H orcuttii using a random Little is known about the basic ecology, coordinate system. We attempted to pair each including population structure and habitat re­ plot containing H orcuttii with a plot lacking H quirements of H orcuttii; however, previous orcuttii, but the spatial distribution of paths and research suggests that H orcuttii is restricted to shape of the vegetation fragments precluded an soils with higher clay content (Oberbauer 1981). adequate paired-design resulting ip unbalanced Research designed to determine the basic ecolog­ replication for H. orcuttii and non-H. orcuttii ical requirements of H orcuttii is needed to plots. conserve this species (Gogol-Prokurat and Os­ Measurements ofplant species abundance were borne 2002). Given the current status of this conducted over a total of 4 sample campaigns species the main objectives of this research were (January and July of 2004 and 2005) in the H to characterize the soil physical and chemical orcuttii plots and non-H. orcuttii plots described properties and the plant community associated above. All individuals rooted within each plot with the extant H orcuttii population at Lux (Chapmari 1976; Barbour et al. 1999) were Canyon. counted and measured for width along 2-axes (the maximum width and the axis perpendicular MATERIALS AND METHODS to the maximum width) and height from the ground surface to the top of the shrub (Bonham Site Description i989). Soil and surface organic matter (litter, which is Field measurements were conducted from dead plant matter > 1 mm in diameter) was January 2004-July 2005 at Lux. Canyon collected in April 2004 to coincide with the (33°1'48"N, 117,0 15'6"W) in the Manchester spring growing season and the main period of Conservation Area in Encinitas, CA. Lux seed germination. Samples were obtained from Canyon is approximately 5 km east of the ocean the non-H orcuttii plots {n = 10) and a subset at an elevation that ranges from 10 m above sea of the H orcuttii plots (n = 10) to preserve as level in the valley bottom to 100 m on the mesa much as possible the paired-sampling design top (Center for Natural Lands Management between H orcuttii and non-H. orcuttii plots. (CNLM) 2005). Vegetation consists of Diegan Surface litter was collected within a 312.5 cm2 sage scrub and southern maritime chaparral rectangular quadrat that was centered on a (CNLM 2005), and the main soil types consist randomly chosen point in each plot. After litter of Altamont clay (Typic chromoxerert) on the removal, soil samples were obtained from mesa top and a loamy, all\l'¥ial Huerhuero surface (0-10 em) and subsurface (30-40 em) complex (Typic natrixeralf) on the eroded slopes soil layers using a 173.5 cm3 bucket auger. Soil and valley bottoms (Bowman 1973). Climate samples were transferred from the core samplers data obtained since 1998 from the National to polyethylene sample bags and immediately Oceanic and Atmospheric Administration for returned to the lab and stored at 4°C until Palomar Airport in Carlsbad, California located analysis. approximately 13 km north of Lux Canyon with similar coastal exposure indicates a maritime Sample Processing and Data Analysis Mediterranean-type climate with warm-dry sum­ mers and cool-wet winters. Average annual Plant species density was quantified as the rainfall is approximately 200 mm (7.9 in.) and number of individuals per species per unit plot average annual maximum and minimum tem­ area and cover was quantified as the area ofeach perature is 19.8 and 12.7°C, respectively. The shrub species per plot divided by the area of the 2009] VOURLITIS ET AL.: HABITAT CHARACTERISTICS OF HAZARDIA ORCUTTII 231 plot. The area (A) of each individual shrub was A S: 16.79'" calculated as rcD2/4, where D is the average D: 1.09 diameter of each individual calculated from the SxD: 0.19 measurements of maximum and perpendicular width. Frequency of occurrence was calculated as the riumber of plots that a particular species was encountered. Indices .of relative abundance were calculated from the estimates of absolute abun­ dance by dividing a given absolute abundance for a particular species by the total abundance of all species. These relative indices were combined to yield an estimate of the index of relative importance (IRI), which was calculated as the 0 200 400 600 800 sum of the individual relative density, cover, and Sand content (mg/g) frequency of occurrence indices (Chapman 1976; Barbour et al. 1999). 8 S: 13.25"' D: 3.68 Soil samples were sieved to remove rocks and SxD: 0.02 organic matter ~2 mm in size prior to laboratory analyses. Litter samples ;were passed through a 1 mm sieve" to remove mineral debris, dried at 70°C for 1 wk, weighed, and ground to pass through a 40 mesh sieve. Total N and P content of soil and litter was quantified using micro~Kjeldahl methods (Bremner 1996). Percent gravimetric soil water was calculated as [(Mr - Mct)l Mct]* 100 where Mr was the fresh mass of soil and Mct was the mass of soil after drying at 0 100 200 300 400 105°C (Robertson et al. 1999). Percent soil Claycontent (mg/g) organic matter was quantified by combusting FIG. 1. The mean (±SE) sand (A) and clay (B) soil at 700°C for 1 h in a muffie furnace (Nelson content of soil in plots where Hazardia orcuttii was and Sommers 1996). Soil bulk density was present (closed bars; n = 10 plots) or absent (open bars; calculated as the mass of dry soil per unit n = 10 plots) at different soil depths. Also shown are volume (Robertson et al. 1999). Soil particle size the results of a 2-way ANOVA (F-statistic) where sub­ distribution was measured using the Bouyoucos site (habitat) and depth were fixed effects (df = 1,36 for hydrometer method (Gee and Bauder 1986). sub-site (S), depth (D), and the sub-site X depth Cluster analysis was used to assess the similar­ interaction (S*D)). *** P < 0.001; ** P < 0.01; * P ity in community composition between the < 0.05. habitats containing and lacking H. orcuttii. For this analysis, the plot values of the IRI of RESULTS each species were used to determine the degree of similarity between discrete plots, and the Soil Physical and Chemical Properties "Euclidian Distance" method was used to determine the relative distance betwt;en each Sand content of surface (0-10 em) and sub­ plot (Hintze 2005). Soil physical and chemical surface (30-40 em) soil in areas lacking H. orcuttii data were analyzed using a 2-way '.analysis of was significantly higher than in areas containing variance (ANOVA) with site (H.. orcuttii VS. H. orcuttii (Fig. lA; F 1,36 = 16.79; P < 0.001). non-H. orcuttii plots) and depth (0-10 and 30­ Clay content of soil in areas with H. orcuttii was 40 em) treated as fixed effects. Data were tested significantly higher than in areas lacking H. for normality and heteroscedasticity using the orcuttii (Fig. lB; F 1,36 = 13.25; P < 0.005). Anderson-Darling and Levene's tests, respec­ Differences in particle size distribution were large tively. Data violating assumptions of normality enough that plots containing H. orcuttii were and heteroscedasticity were LN-transfon:ned to characterized as having a sandy clay loam soil fulfill the assumptions of ANOVA. Cluster while plots lacking H. orcuttii were characterized analysis and ANOVA were performed using as having a sandy loam soil. Plots containing H. NCSS (version 2004, Kaysville, Utah, USA). orcuttii also had significantly· higher soil water Differences in litter pool biomass and N and P content (Fig. 28-), especially in the sub-surface, content between H. orcuttii and non-H. orcuttii and a significantly higher soil organic matter plots were analyzed with a randomized-t-test (SOM) ¢onten:t (Fig. 2B). Total soil N content (Sokal and Rohlf 1995) using MS-Excel (Chris­ (Fig. 3~ was significantly higher in plots con­ tie 2004). taini:gg/'H orcuttii, especially in the surface soil :~ ..., ·• . ~ .. ·.~ 232 MADRONO [Vol. 56

A A S: 8.68** 0: 46.99*** SxD: 0.08

0-10 0-10

30-40 30-40

0 50 100 150 200 250 I 0 20 40 60 80 100 Soil water content (mg/g) .s:: Total soil N(gN/m2) i5. CD Cl B S: 1.38 B S: 10.88** D: 1.68 D: 1.42 SxD: 2.11 SxD: 1.78

0-10 0-10

30-40 30-40

0 6 8 10 12 14 16 18 0 10 20 30 40 50 60 2 4 Soil organic matter content (mg/g) Total soil P (gP/m2) FIG. 2. The mean (±SE) water (A) and organic matter FIG. 3. Mean (±SE) total soil N (A) and P (B) content (B) content of soil in plots where Hazardia orcuttii was in plots where Hazardia orcuttii was present (closed ·present (closed bars; n = 10 plots) or absent (open bars; bars; n = 10 plots) or absent (open bars; n = 10 plots) n = 10 plots) at different soil depths. Also shown are at different soil depths. Also shown are the results of a the results of a 2-way ANOVA (£-statistic) where sub­ 2-way ANOVA (£-statistic) where sub-site (habitat) site (habitat) and depth were fixed effects (df = 1,36 for and depth were fixed effects (df = 1,36 for sub-site (S), sub-site (S), depth (D), and the sub-site X depth depth (D), and the sub-site X depth interaction (S*D)). interaction (S*D)). ***P < 0.001; **P < 0.01; *P ***P < 0.001; **P < 0.01; *P < 0.05. < 0.05. plots lacking H. orcuttii ("B" plots; Fig. 5), while layer. In contrast, total soil phosphorus (P) the second major grouping consisted of the plots content was similar for plots with and without where H. orcuttii was present ("A" plots; Fig. 5). H. orcuttii (Fig. 3B). This clustering described the maximum variation The N and P concentration ofsurface litter was in the sample plots (r = 0.89), and indicated two not statistically different between H. orcuttii and discrete sub-communities at Lux Canyon. non-H. orcuttii habitats; however, because sur­ Some species were common to both sub­ face litter biomass was nearly 3-ti.j:nes higher in communities, including Adenostoma fasciculatum, non-H. orcuttii plots (P = O.OQ..3; ::Ft,g. 4A), litter Artemisia californica, Eriogonum jasciculatum N and P pools sizes were significantly higher in Benth., Quercus dumosa Nutt., Rhus integrifolia, non-H. orcuttii plots (Fig. 4B, C). Thus, while H. and Deinandra fasciculata (DC.) Greene (Ta­ orcuttii plots had significantly higher SOM and ble 1). However, A. fasciculatum, Q. dumosa, and total soil N, plots lacking H. orcuttii had a R. integrifolia had higher mean values of IRI over significantly higher surface litter pool and litter N the 2-year field study in non-H. orcuttii plots, content. while E. fasciculatum, Dudleya edulis (Nutt.) Moran, and D. fasciculata were more abundant Community Composition in H. .orcuttii plots (Table 1). Other species including Mimulus aurantiacus Curtis, Xylococcus Cluster analysis of the sample plots indicated bicolor Nutt.,: Yucca schidigera Ortgies, and Y two discrete vegetation assemblages (Fig. 5) whipplei Tor.r. were conspicuously lacking in H. based on the importance values\(JRI) of the plant orcuttii plots; ~hile Lotus scoparius (Torr. & A. species. The first group consisted of all of the Gray).Otdey, }

2000 A Gray) Britton & Rose, and Baccharis pilularis DC. were observed only in H orcuttii plots "' E"' 1500 (Table 1). .QN Significant differences in total density and ~E g 0,1000 cover were also apparent between H. orcuttii a.~ 2 and non-H orcuttii plots (Fig. 6). Plots lacking 5 500 H orcuttii had significantly higher cover in 2004 (Fig. 6A), which explains in part the significantly 0 higher surface litter pool (Fig. 4A). Plots lacking 25 B c H orcuttii also had significantly lower density ~ 20 than H. orcuttii plots during each measurement 8N campaign (Fig. 6B) suggesting that, on average, zJ2 15 a :a, non-H orcuttii plots were dominated by fewer e.~ 10 but larger shrubs. One of the most abundant ,gJ shrubs in plots lacking H orcuttii was A. :::; 5 fasciculatum (Table 1), which is a chaparral shrub 0 that can reach heights of 2m and an area of 3­ c 4 m 2 (Munz 1974; Riggan et al. 1988). c 0.8 c(]) Temporal variations in stand cover and density 8N" 0.6 were relatively higher in H orcuttii plots thh in a.J2 non-H orcuttii plots. For example, H orcuttii 0~ 0.4 e.~ plots experienced a nearly 4-fold increase in stand ,gJ cover (Fig. 6A) and a more than 5-fold increase :::; 0.2 in stand density (Fig. 6B) between July 2004 and 0.0 January 2005. Over the same period plots lacking Present Absent H orcuttii experienced no increase in stand cover Hazardia orcuttii (Fig. 6A) and a 2-fold increase in stand density FIG. 4. Mean (±SE) litter biom!l-SS (A) and total litter (Fig. 6B). Thus, H orcuttii plots experienced N (B) and P (C) pool sizes in plots where Hazardia larger temporal variation in overall plant species orcuttii was present or absent (n = 10 plots per sub­ abundance than non-H. orcuttii plots, which may site). Also shown are the results of a randomized t-test, have implications for H. orcuttii recruitment and where the probability of committing a type-I error (P­ survival. value) was calculated over 1000 iterations. Temporal variation in the index of relative importance (IRI) of the six dominant shrub species (excluding H orcuttiz) present in both

PIDI .------85 .------82 .------68 L------66 .------67 .------610 Non-H. orcuttiiplots .------69 L------84 .------83 L------61 .------~'~··------A1 L------~~------A9 .---~~------AS ~~------As .------~----A7 r------A4 .------A6 H. orcuttffplots .------A1D .------A13 L------A12 r------A2 .------A3 ~------A1 iilliiliiillilillilillllljlllililiilililllllllllililiijjiijiillllliiliil)llliiliillilliil 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 Dissimilarity

FIG. 5. Cluster analysis of the index of relative importance. (II#!J values of plant species observed in plots containing (Al-13) and lacking (Bl-10) Hazardia orcuttii. ._.; ·• ,C 234 MADRONO [Vol. 56

TABLE 1. MEAN (±lSD) INDEX OF RELATIVE IMPORTANCE (JRI) FOR ALL SPECIES OBSERVED IN 12.56 M2 PLOTS CONTAINING (PRESENT; N = 13) AND LACKING (ABSENT; N = 10) HAZARD/A ORCUTT!! AT LUX CANYON, ENCINITAS, CALIFORNIA. Data were collected over 4 sampling campaigns conducted in January and July of 2004 an<1 2005. IRI was multiplied by 100; - not observed. Nomenclature and habitat data according to Hickman (1993).

Hazardia orcuttii Species Present Absent Adenostoma fasciculatum Hook. & Am. 28.9 ± 23.3 66.8 ± 9.8 Artemisia californica Less. 64.6 ± 36.4 49.9 ± 8.9 Dudleya edulis (Nutt.) Moran 13.8 ± 5.2 5.3 ± 3.2 Eriogonum fasciculatum Benth. 42.7 ± 22.1 14.0 ± 3.7 Hazardia orcuttii (A. Gray) Greene 32.4 ± 15.5 Ferocactus viridescens (Torr. & A. Gray) Britton & Rose 7.3 ± 2.0 ' Mimulu's aurantiacus Curtis 37.6 ± 4.1 Opuntia littoralis (Engelrn.) Cockerell 1.5 ± 1.1 5.5 ± 0.5 Quercus dianosa Nutt. 2.2 ± 1.1 17.0 ± 4.5 Rhus integrifolia (Nutt.) W. H. Brewer & S.cWatson 13.4 ± 7.6 47.6 ± 18.7 Xylococcus bicolor Nutt. 2.8 ± 0.7 Yucca schidigera Ortgies 8.6 ± 4;1 Yucca whipplei Torr. 2.8 ± 1.9 Lotus scoparius (Torr. & A. Gray) Ottley 8.2 ± 3.4 Deinandrafasciculata (DC.) Greene 55.6 ± 79.2 9.9 ± 12.5 Marah macrocarpus (Greene) Greene 3.9 ± 5.4 Dichondra occidentalis House 1.1 ± 2.2 Centaurium venustum (A. Gray) B. L. Rob. 1.7 ± 3.4 Anagallis arvensis L. 0.3 ± 0.6 0.6 ± 1.3 Cholorgalum parviflorum Wats. 5.8 ± ll.5 10.9 ± 21.8 Zigadenus fremontii (Torr.) S. Watson 0.4 ± 0.9 Dichelostemma pulchella (Salis b.) Heller 0.9 ± 1.7 0.5 ± 1.0 Baccharis pilularis DC. 0.4 ± 0.8 Eriophyllum confertiflorum (DC.) A. Gray 0.6 ± 1.2 Gnaphalium californicum DC. 0.4 ± 0.7 1.8 ± 3.7 Gnaphalium sp. 0.6 ± 1.3 0.6 ± 1.2 Ciyptantha sp. 0.5 ± 1.1 Stephanomeria sp. 0.5 ± 0.9 Thistle 8.6 ± 16.0 5.8 ± 7.0 Unknown annual 0.9 ± 1.7 1.2 ± 1.4 Unknown herbaceous perennial 0.6 ± 1.2 Unknown grass 10.7 ± 18.7 2.3 ± 3.3 plot types revealed substantial changes in shrub potentially key difference was soil texture, where species composition over 2004-2005, especially in H. orcuttii plots had significantly higher clay the H. orcuttii plots (Fig. 7). In H. orcuttii plots, content and lower sand content than non-H. A. fasciculatum and D. fasciculata exhibited 3-4 orcuttii plots (Fig. 1). This observation is sup­ fold increases in IRI during the study period, ported by previous research and is consistent with while 1· californica and E. fascic;'!-latum experi­ the notion that H. orcuttii may be a soil endemic enced a 2-3-fold decline in, IRI (Fig. 7). In (Oberbauer 1993). Gravimetric soil water content contrast, plots lacking H. orcuttii experienced was significantly higher in H. orcuttii plots substantially less temporal ,yariation in relative (Fig. 2A), which presumably reflects the higher abundance; however, Q... : !ntegrifolia was an clay content of the soil. Soil texture controls a obvious exception (Fig. 7). These data indicate variety of processes that control plant species rapid and dynamic specil'ts turnover in H. orcuttii distribution, including soil water holding capac­ plots and more stable community dynamics in ity, nutrient retention, organic matter stabiliza­ non-H. orcuttii plots. tion, seed germination, and seedling recruitment (Baskin and Baskin 1990; Oberbauer 1993; DISCUSSION Schimel et al. 1985; Kluse and Doak 1999; Walck et al. 1999; Hook and Burke 2000). However, it is Soil Physical and Chemical Properties unclear why H. orcuttii at Lux Canyon is restricted ,to soil with higher clay content. For Some soil physical and chemical properties at examp~~' percent germination of H. orcuttii seeds Lux Canyon were significantly different between was a~iually higher in soil types that had lower plots containing and lacking H. orcuttii. One ctay~,s~ntent (Miller 2008); thus, H. orcuttii is ·.~:;·:· .2009] VOURqTIS ET AL.: HABITAT CHARACTERISTICS OF HAZARDIA ORCUTT!! 235 apparently not restricted to soil with higher clay 3.5 A content because of seed germination. ...._ Present m3.o ...0 .. Absent Hazardia orcuttii plots had higher soil organic ·~ matter (SOM) content than non-H orcuttii plots 1!"' 2.5 :::> (Fig. 2B), which presumably indicates differences ~ 0 in plant species composition and/or rooting 81:~a 2o. depth between habitats (Jobbagy and Jackson ~-m 1.5 2000). The increase in SOM was apparently not 1- iii due to an increase in aboveground litter input 'E 1.0 because the surface litter pool was nearly 3-fold c."' ?··'""' ·? .. ··' '' lower in H. orcuttii plots (Fig. 4A). Rather, H 'l 0.5 orcuttii plots had a higher abundance of shrubs associated with coastal sage scrub (Table 1), 0.0 B which typically have a shallower, more horizo~­ 14 tally-distributed root system than species charac­ 12 teristic of evergreen chaparral (Hellmers et al. "'~ "C"' 1955). Hazardia orcuttii plots also had higher >,C: 10 total soil N, suggesting higher ovet311 fertility ~e 8 compared to non-H orcuttii plots (Marion and "ON~ "' -ro-E Black 1988). Given that SOM represents a large -o-(}) 6 pool of N in terrestrial soils (Hook .and Burke 1-gj ~ 4 2000), these results presumably reflect the signif­ "C 5 icantly higher SOM content observed in H 2 orcuttii plots. Surface litter was more than 3-times higher in 0 * ** * ** non-H. orcuttii plots, which has important Jan04 Jui04 Jan05 Jul05 implications for the germination and recruitment Measurement period of H. orcuttii. For example, the germination and FIG. 6. Average (±SE) total shrub cover (A) and seedling recruitment of Chorizanthe pungens density (B) in plots containing Hazardia orcuttii (closed­ Benth. var. hartwegiana Reveal and Hardham circles, solid-lines; n = 13 plots) and plots lacking H. (Polygonaceae) is reportedly inhibited by chap­ orcuttii (open-circles, dotted-lines; n = 10 plots). arral vegetation, possibly as a result of allelopa­ Differences in mean cover and density between habitat thy and/or the development of a larger surface types were assessed using a two-sample t-test. * P < 0.05; ** p < 0.01. litter pool that· alters the microclimate and reduces light availability at the soil surface (Kluse and Doak 1999). Results from germination soil texture, which exerts a strong influence on experiments indicate that percent germination of soil water availability and plant species distribu­ H. orcuttii seeds was significantly lower in tion (Westman 1981). Hazardia orcuttii plots had complete darkness,. which may simulate light significantly lower cover and higher density conditions under a deep surface litter layer (Fig. 6) and more rapid and dynamic species (Miller 2008). While there are substantial turnover (Fig. 7) than plots lacking H. orcuttii, seasonal variations in the production of above­ and it is possible that these interspecific dynamics ground litter and the size of the surface litter affect H. orcuttii recruitment, survival, and pool in coastal sage scrub and chaparral fecundity. For example, the higher temporal (Vourlitis et al. 2009), the difference in the litter variation in species composition in H. orcuttii pool between plots containing:'and lacking H plots implies higher variation in the intensity of orcuttii observed in April coincides with the most competitive interactions, availability of "safe active time for seed germinati~ for chaparral sites" important for If. orcuttii recruitment, and coastal sage shrubs. T~e ·relatively larger and/or available resources (Menges 1990; Watson surface litter pool observed in non-H. orcuttii et al. 1994; Kluse and Doak 1999; Walck et al. plots may inhibit seed germination and/or 1999). Similar interspecific controls on plant seedling recruitment, which may be one impor­ growth, fecundity, and recruitment have been tant mechanism causing H. orcuttii to be observed for other Asteraceae including Coreop­ restricted to more clay-rich soils. sis lanceolata L. (Folgate and Scheiner 1992), Ratibida columnifera (Nutt.) Wooton & Standi. Community Composition (Vargas-Mendoza and Fowler 1998), Solidago shortii Torr; & A. Gray (Walck et al. 1999) and Our results indicate fundamental differences in Deinandra conjugens (D. D. Keck) B. G. Baldwin species composition between H. orcuttii and non­ (Bauder et l'Jil. 2002). Presumably similar process­ H. orcuttii plots (Fig. 5). These differences are es may b~ important in limiting the local presumably due in part to spatial variations in · distributipifof H. orcuttii. . ... ,:. ~.~ 236 MADRONO [Vol. 56

Af graphic range, and narrow habitat tolerance 0.8 0.8 suggest that H. orcuttii is rare by all criteria. 0 Protection of Lux Canyon from human 0.6 0.6 0· ·······o. .0 degradation and fire should be a high priority. 0.4 0.4 The Lux Canyon site is used for recreation purposes, and human activities lead to the 0.2 0.2 creation of paths, trampling and damage of ~ 0.0 0.0 vegetation, accumulation of waste, and urban ~ J\_._Present 0.8 Ef 0.8 Ri runoff. These threats increase the potential for 8 ...o .. Absent s:: 0. catastrophic fire, which can either damage or ~ 0 0.6 0.6 a. ·o completely eliminate that only known U. S. .5 population ofH. orcuttii. ·o ...... ·0 ~ ACKNOWLEDGMENTS · ~ 0.2 0.2 0 o~...... o...... o...... o X This research was supported through a grant

Book Series No. 5, Soil Science Society ofAmerica, ---. 1993. Soils and plants oflirnited distribution in Inc., Madison, WI. the Peninsular Ranges. Frem