- Social Sciences
- African Studies
- American Studies
- Asian Studies
- Communication Sciences
- Ethnic Studies
- European Studies
- Gender Studies
- Physical Sciences
- Life Sciences
- Animal Communications
- Cell Biology
- Evolutionary Biology
- Food Science and Technology
- Human Anatomy
Patterns of Allozyme Diversity in the Threatened Plant Erigeron parishii (Asteraceae)
by Maile C. Neel, Norman C. Ellstrand
Patterns of Allozyme Diversity in the Threatened Plant Erigeron parishii (Asteraceae)
Maile C. Neel, Norman C. Ellstrand
American Journal of Botany
Updated: January 24th, 2013
American Journal of Botany 88(5) 810-818 2001
THREATENED PLANT ERZGERON
MAILEC. NEEL~AND NORMANC. ELLSTRAND
Department of Botany and Plant Sciences and Center for Conservation Biology, University of California,
Riverside. California 92521-01 24 USA
Thirty-one occurrences of Erigeron pnrisizii, a narrowly endemic plant threatened by mining, were sampled for allozyme diversity. This taxon held considerable genetic variation at the 14 allozyme loci surveyed. Species (e.,o., alleles per locus [A] = 4.3 and proportion of polymorphic loci [PI = 0.64) and population (e.g., A = 2.15 [SD = 0.31 and P = 0.53 [SD = 0.121) genetic diversity measures were higher than expected for nan.owly ende~nic plant taxa. The proportion of polymorphic loci and numbers of alleles per locus indicated that E. pnrishii has not experienced severe or long-lasting population bottlenecks. Within-population f indicated low to moderate levels of inbreeding. Populations were only ~iioderately differentiated (theta-p = 0.12), suggesting either that there is sub- stantial gene flow ainong populations or that populations have not been isolated long enough to detect effects of genetic drift. There was no significant differentiation ainong popnlations in different vegetation types nor was there a relationship between genetic distance and geographic distance among sites. Continued fragmentation by mining activities would isolate populations, disrupting gene flow, exacerbating loss of diversity, and increasing inbreeding in the remaining fragments. Protection of large, interconnected populations throughout the range of the taxon is warsanted to maintain processes that have sustained the observed diversity.
Key words: Erig~rn~i
parishii; genetic diversity: plant conservation; reserve design
The major cause of extinctions for both plants and animals is loss, degradation, and subsequent fragmentation of once-continuous habitat (Holsinger and Gottlieb, 1991; Falk, 1992; Clegg et al.. 1995; Wilcovc et al.. 1998). Fragmentation, a byproduct of habitat destruction, reduces habitat patch size, increases edge-to-area ratio, increases distance among patches and increases resistance to migration among patches (Opdam et al., 1994). Fragmentation can also rapidly alter interactions among populations within species, among species, and be- tween species and abiotic features of the environment (Coop- errider, 1991; Thompson. 1996). Such changes can increase extinction risks from-demographic and genetic stochasticity by decreasing effective population sizes and illcreasing rates of inbreeding and genetic drift (e.g., Barrett and Kohn, 1991; Ellstrand and Elam, 1993; Young, Boyle, and Brown, 1996). At the same time, rates of gene flow can either increase or decrease, depending on how dispersal patterns are affected by fragmentation (e.g., Ellstrand and Elam, 1993). Most frequent- ly, gene flow is assumed to decrease as distances between remaining populations increase. Under conditions causing loss of genetic variation and decreased gene flow, dccreased het- erozygosity and increased differentiation among populations are expected. Such patterns are thought to be detrimental and to increase extinction risks. Rivks are highest when historically
I Manuscript received l l April 2000; revision accepted 18 July 2000.
The authors thank J. Frie (US. Fish and Wildlife Service, permit number PRT 8265 IS), W Bretz (Burns Pinyon Ridge Reserve, University of Califor nia Reserve System), D. Volgarino (USDA Forest Service), and T Egan (Bu- reau of Land blanagement) for facilitating permits for tissue collections. L. Blancas, M. Cummings. B. Henderson, and S. Meyer, assisted in the field. L. Bhattia, L. Blancas. and J. Clegg provided valuable assistance in the lab. E. Elle, R. Farr. L. McInerney, and two anonymous reviewers provided valuable comments on the manuscript. This work was partially funded by a Switzer Environmental Fellowship, by the University of California Riverside, and by the U.S. Environmental Protection Agency Project Number R826102-01-0.
'Author for correspondence, current address: Josephine Bay Paul Center for Conlparative Molecular Biology and Evolution, The Marine Biological Laboratory. Woods Hole, Massachusetts 02543 USA (e-mail: mailen@mail. ucr.edu; telephone 508-548-3705 ext. 6634; fax 508-457-4727).
large populations with substantial interpopulation gene flow are reduced in size and isolated from one another (Bal-sett and Kohn. 1991).
A major objective of conservatiorl biology is to maintain biodiversity by promoting persistence of species in their native ecosystems over time (Harrison. Miller, and McNeely, 1984; Falk, 1992). Because chances of persistence decrease in de- graded and fragmented habitat, much attention has focused on protecting areas from destructiorl to slow habitat loss and frag- mentation, and on managing those areas to preserve and en- hance biodiversity (e.g., Noss. O'Connell, and Murphy, 1997). Such areas go by many different names; we use the term "re- serve" here. Many reserves are designed and established to conserve s~ecific rare taxa. which often occur in small. iso- lated popuiations or fragmented populations that have been reduced in size. A common goal of such reserves is to support large enough populations to prevent inbreeding and genetic drift. This goal is most often met by protecting sufficient hab- itat rather than directly focusing on genetic diversity present within the target species. It is assumed that if sufficient habitat is maintained to protect against environmental stochasticity, loss of genetic diversity is not an immediate concern (Tem- pleton et al., 1990; Schemske et al.. 1994: Gaines et al., 1997). However, because genetic diversity contributes to species per- sistence, its direct measurement can be an important priority.
Quantifying the organizatioil of genetic variation over pop- ulations of a sensitive species call help in prioriti7illg sites and management choices for maintaining that variation. For ex- ample, highly diverse or differentiated populations could be targeted for protection while depauperate populatiolls might betargeted for management actions to restore diversity (e.g.. Godt, Johnson, and Hamrick, 1996; Petit, El Mousadik, and Pons, 1998). Information on genetic diversity patterns also provides insight into evolutionary and demographic history of a taxon (Milligan, Leebens-Mack, and Strand, 1994). Under- standing the relative importance of processes that structure di- hersity within and among populations (specifically inbreeding, gene flow, genetic drift, and selection) call provide both a
Fig. I. General locatiolls of sat~ipling sites for El.igerot7 pnri.rhii shown in context of the state of California and the boundary of the Sari Bernardino National Forest. The sanlple site labeled EPBPRR was located at the Burns Pinyon Ridge Reserve and was disjunct from the remaining sites.
means to assess future risk of erosion of diversity and a means for designing effective conservation strategies for rare taxa. For example, if genetic diversity is primarily held within pop- ulations, fewer populations might need to be conserved to rep- resent the range of variation within a taxon. In such cases genetic diversity might be less important a criterion for se- lecting particular populations than other criteria. Alternatively, a taxon with most of its variation partitioned among popula- tions would require protection of a larger proportion of exist- ing populations to maintain variation present in the taxon. The structure of genetic diversity within and among populations also has important implications for developing sampling strat- egies for restoration and reintroductions (e.g., Brown, 1989; Ceska, Affolter, and Hamrick, 1997). Finally, genetic diversity information can guide future research by focusing attention towards areas of particular potential concern. For example, if genetic diversity is high, and populations do not appear to be at risk of losing that diversity, research efforts can be focused on other ecological characteristics related to survival of indi- viduals in relation to threats faced by a taxon.
One example of a plant species that has been subjected to fragmentation for decades is the federally threatened Erigeron pnrishii A. Gray. This taxon is primarily endemic to limestone and dolomite substrates (collectively known as carbonate sub- strates) in the San Bernardino Mountains. These deposits pro- vide ore that is extremely valuable in construction, industrial, and pharmaceutical applications (Calzia, 1993). Consequently,
E. parishii has suffered substantial fragmentation by limestone mining activities (Skinner and Pavlik, 1994; USFWS, 1994).
Conservation planning efforts for this taxon are focusing on establishing some areas that will be protected from mining while allowing mining at other sites (USFWS, 1997). The pur- pose of our research was to characterize the intra- and inter- populatio~l genetic structure of E. yarishii. This information can provide guidance in overall reserve design and contribute to ranking the conservation significance of individual popu- lations.
MATERIALS AND METHODS
Etigerotl pcrrisilii is a low-growing, perennial herb (Hickman, 1993). Individual plants grow from a taproot and typically reach 1-3 dnl in height. The 1-3 ~1x1long, narrow leaves emerge in the spring and become dry by summer's end, at which time the stems remain dormant above ground. Most flowering occurs between May and July, but some individuals bloom into August, especially after abundant summer rains. Lavender ray flowers and yellow disk flowcrs occur in capitula borne at the ends of leafy stems. Indi- vidual plants typically have iiiany capitula and each capitululn produces hun- dreds of small achenes (2-3 X 0.5 rnni). The pappus on each achene is only 2-3 mm and is unlikely to facilitate long distance dispessal. Nothing is known about E. l~nrishii's mating system; everything from obligate, insect-mediated outcrossilig to agarnospermy is known from the genus (Moldenke, 1976; Rich- ards, 1990). In addition, tnatiy mernbers of the genus Ei.iger.oi7 are clonal. Mating system and clonality were not specifically exnniilied in this study; however, some general inferences can be made from our results.
As mentioned above. E. parislzii is primarily endemic to carbonate sub- strates in the San Hernardino Mou~itailis (Fig. 1) (Skinner and Pavlik, 1994; USFWS, 1994). These carbonate substrates occur as varying sized outcrops and alluvial deposits totaling -13 200 ha. E. porishii most often occupies ailwium along ephemeral drainages but is also found on rocky slopes. Oc- currences are patchily distributcd on apprnpriate substrate in creosote bush- bursage scrub, black bush scrub, and various phases of pinyon-juniper woodland between -1 200 and 2000 ni elevation. Two occurrences on quartz mon- zonite at the eastern end of the mountain range and oiie historic location in the Little San Ber~~ardino Moiintains are exceptions to strict endemism to carbonate substrates in the San Bernardino 3lountains. Known occusrences in July 1999 totaled -453 ha (S. Redar, personal cornn~unication, San Bernar- dino National Forest).
Where found, E:'.l~a~ishii
populations can be quite large, comprising thou- sands of indivitluals, though they arr usually smaller. In fact, this taxon has more extant populations and a larger nuniber of total individuals than the typical endangered plant species (Holsinger and Gottlieb, 1991: Ellstrand and Elam, 1993). Existing populations arc not apparently in demographic decline. Sizes of some populations are above those at which inbreeding and genetic drift are typically of irnmediatc concern (Simberlofl', 1988: Ellstrand and Elam, 1993). The main concern for persistence of E. pnrislzii is continued habitat destruction and fragmentation, primarily from limestone mining activ- ities. Populations of and habitat for E, porishii have been lost to mining, and there is inlinineut threat of continued loss (USFWS, 1094).
Most E. pnrishii populations arc threatened because they exist primarily on private land or on public land vith no protective status. Additionally, most of E. parishil's range osi public Innti is ~inder valid mining claim and thus at risk from future mining activities. As a result of existing habitat and popu- lation losses. continuing threats, and lack of protectilfc n~echanisn~s,
the U.S. Fish at~d Wildlile Service listed this taxon as threatened (USRVS, 1994). Efforts to conserve this species involve designing a rcscrve system that will be protected from lnining as well as restoration of degraded sites. Due to the high value of the ~uineral ore, not all existing populations \\.ill be maintained. Information on patterns of genetic diversity can assist in evaluation of pop- ulations for protection.
Sampling-Thirty-one occurrences of E. pari.siiii were sampled for allo- zyme diversity (Fig. 1). Sampling sites mere selected stratified randomly to represent the ecological range of the taxon as follows. Twenty-five allozyrne sampling sites coincided with plot locations for a related study in which veg- etation conlrriunity conlposition xas described (Neel, 2000). Sixtl-one of the 669 0.04-ha vegetation plots in that study included E. pnrishii. Twenty of these plots occurretl in black bush scr~tb (BBS): 27 in singleleaf pinyon-Utah juniper woodlands IPJ); and 14 in singleleaf pinyon woodlands (PY). Allo- zyme sampling sites were chosen from these plots roughly in proportion to the number of plots within each vegetation type. Six sample sites not asso- ciated with vegetation plots were included to represent ecological and geo- graphical variation not included in those plots. One of these sites was a tlis- junct occurrence of E. (~ari.rlzii (EPHPRR) at the Burns Pinyon Ridpe Resene (University of California Natural Reser\:e System) that represented the east- ernmost extant populations (Fig. I). An historic collection further east from the Little San Bernardiiio Mountains has not been located recently. The Burns Pinyon Ridse Reserve occurrence was found on quart7-mouzonite in pinyoii- California juniper woodland (PJUCA). l'he substrate was atypical for this taxon, and the vegetation type did not support E. pnrisizii at any other location away from this area. Site EPD135CH was located at the lower elevational limit of the taxun in Creosote Hush-Bunobush Scrub (LATR) vegetation on carbonate alluvium. Site EPBCQ was located near the western edge of the range where populations were not often found in plots. Vegetation composi- tion from the three remaining sites not associated \vith vegetation plots was subsequently sampled by San Bernxdino National Forest botanists (D. Vol- gnrino, personal communication, Sail Bernardino National Forest). One of these plots was in creosote bush-white bursage scrub and the other two were in the single-leaf pinyon woodland. Densities of E, ytrrirhii were crudely estimated using counts from the 0.04-ha vegetation plols where available.
Up to 200 mg of leaf tissue wcre collected from 30 indi~iduals per sample site where possible (a total of 032 individuals were sampled) (Table I). We were unable to assess genotypes for all loci for all individuals and thus sample sizes averaged over loci are given for each site (Table 1); sample site EPB203 was particularly problematic. Additionally, we were able to obtain allozyme data from only 28 or 29 of 30 individuals at three sites. One other site (EPD135CB) supported only seven individuals, fi~e of which hati lice leaf material available for sampling. This site was retained in analyses despite the small sample size because the sarnple represented most of the extant popu- lation and the site represented the lower elevation:il limit of the taxon. Within sites, individuals were chosen from throughout the population within the gen- eral vicinity of the vegetation plot. In general. individuals sampled within sites were separated by 5-30 m. Leaves were trancported on ice to the lab- oratory, where they were rel'rigerated at 8°C.
Allozyme electrophoresis-All samples were initially run from freshly ex- tracted tissue within 7 d of collection. Approximately 30 mg of' leaf tissue was ground in five drops of a 0.1 moliL, pH 7.5 tns-HC1 buffer, Each milliliter of buffer contained 0.34 mg ethylenediamirie tetracetic acid (disodium salt).
0.75 mg KCI, I ~1 mercaptoethanol, 2.03 rng hlgC1, and 30 mg polyvinyl- pyrrolidine-40. This extract was soaked onto chromatography paper wicks that were loaded directly into 9% starch gels. Extra wicks made from these ex- tractions were frozen at -80°C for use in contir~natory runs.
The follow in^ enzyme systems were assayed in all individuals in all pop- ulations: acoilitase (XCO), aspartate ailiino transferase (.4AT), fl~iorescent esterase (FEST). leucine amino peptidase (LAP), 6-phosphogluconate dehydro- genase (6PGD). phosphoglucomutase (PGMI, shikimate dehydrogenase (SKDH), triose phosphate isomerase (TPI), and uridine diphosphoglucose py' ophosphorylase (UDP). Aconitase. SKDH, and 6PGD were resolved in a pH
7. continuous morpholine-citrate pel system run at 30-35 mA and 150 V Lbr 6 11 (O'l/lalley. Wlieele~; and Guries, 1980). Pliosphogluconiutase and TPI wcre resolved on a discontinuous lithii~m hydroxide-borate gel system at 75 mA and 200 V for 5.5 h (Heywood. 1980). Lcucinc amino peptitlase was resolved on a tris-EDTA-horate gel system run at 50 mA and 200 V for 5.5 h (Heywood, 1980). Aspartate amino transferase and UDP were resol'red in both lithium hydroxide and tris-EDTA-borate gel systems.
The genetic basis of allozymes is well known ibr angiosperlns in general and specifically for inany species in the family Asteraceae (Weden and Wen- del, 1989). Our interpretations of the banding patterns observed were bascd on those previous studies and on kno~~n
quaternary structure of thc enzynics studied. Numbers of loci scored for each enzyme system were as follo\cs: ACO (3), AAT (I), FEST (2),LAP (I), 6PGD (2), PGhl (I), SKDH (1), TPI (2), UDP (I).
Data analysis-Total numbel of alleles (TA), TA 5tandard17eJ by sample size, proportion of polymorphic loci (P), number of alleles per locus (A). number of alleles per polymorphic locus (A;,). observed heterozygosity (H,), and expectztl heterozygosity (He) were calculated for the whole species and for each sampling site using the computer program GDA (Lewis and Zaykin, 2000). A locus was considered polymorphic if the most common allele had a frequency <0.95. The number of private alleles within sites (that is, alleles detected at only oiie site) was also counted using GDA.
One sarnple (EPD13SCB) was much smaller than all others (N =: 5 vs. 30). Thc population represented by this sample was also by far the smallest en- countered (h' = 7 vs. hundreds or thousands). Differences in both population and sample size confourld direct conlparison of genetic diversity in this sam- ple site with other sites. To more directly co~llpare levels of genetic diversity we controlled for differences in sample size by randomly choosing five in- dividuals from each site aild calculating all genetic diversity sumrliary statis- tics on this reduced data set. This comparison allowed LISto determine whether the observed levels of diversity in the sillall population were cornparable to those found in equivalent santples from larger populations.
The proportion of individuals with unique genotypes and the number of individunls sharing a genotype in each sample site wcre calculated to assess the potential for agamosperiliy and vegetativc reproduction (Ellstrand and Roose, 1987). Only individuals with data from at least ten loci were incl~~ded in this analysis (a total of five iildividuals were exclutled). Furthermore, site EPB203 was not included because of the large number of individuals from which ue could score only partial genotypes. Thus this analysis was based on 868 individuals from 30 sites.
Hierarchical population genetic structure was described using j'(F,,), F
May 200 11 NEELAND ELLSTRAND-GENETIC PARISH11 813
DIVERSITY IN ERIGERO~V
TABLE1. Sample sizes averaged over loci (N), proportion of polymorphic loci (P), alleles per locus (A), alleles per polymorphic locus (A,),
expected heterozygosity (H,), observed heterozygosity (H,), an estimate of the fixation index over all loci in the population Cf), total number
of alleles (TA), TA standardized by sample size (Stand. TA), number of private alleles, and mean genetic distance to all other populations for
14 allozyme loci in 31 populations of Erigeron pnrishii.Vegetation types in which each sample site occurred are indicated; sample abbreviations
are as in the text (Materials and methods: Sampling). Plant density is the number of individuals in one, 0.04-ha plot. The five highest ranking
values for standardized TA, P, H,, and genetic distance, the five lowest ranking values forf, and all occurrences of private alleles are in boldface
Prlvz~te Mean Sample N P A AD H, H,, f TA Stand TA alleles dlrtance Vegetation type Plant den~ity
EPB26 28.6 0.64 2.4 2.9 0.20 0.18 0.12 33 1.15 0 0.02 PY EPB3 1 29.9 0.71 2.4 3.0 0.20 0.20 0.02 34 1.14 0 0.04 PY EPB54 29.9 0.31 1.5 2.5 0.17 0.17 0.00 20 .67 0 0.06 PJ EPB55 29.9 0.43 1.8 2.8 0.18 0.17 0.04 25 .84 0 0.04 PJ EPB202 29.9 0.36 2.0 3.2 0.15 0.14 0.08 28 .94 0 0.09 PY EPB203 50.6 0.57 2.7 3.6 0.15 0.13 0.13 38 .75 1 0.02 PY EPCS 28.9 0.50 2.0 2.9 0.17 0.16 0.05 28 .97 0 0.03 BBS EPC 13 29.9 0.57 2.4 3.5 0.23 0.20 0.11 34 1.14 2 0.03 BBS EPC 19 29.5 0.36 2.2 3.4 0.13 0.14 -0.09 30 1.02 0 0.04 PJ EPC 102 28.5 0.50 2.2 3.1 0.15 0.14 0.05 31 1.09 0 0.01 LATR EPC 104 27.5 0.50 2.1 3.0 0.14 0.13 0.06 29 1.05 0 0.02 BBS EPDS 29.6 0.50 1.9 2.9 0.14 0.14 -0.01 27 .91 0 0.02 PJ EPD6 29.0 0.57 2.4 3.0 0.18 0.17 0.09 34 1.17 0 0.02 PY EPD 12 27.6 0.71 2.6 3.1 0.22 0.20 0.07 36 1.30 1 0.03 BBS EPDlOO 28.7 0.29 1.9 3.0 0.14 0.13 0.07 26 .91 0 0.03 PY EPD102 29.7 0.36 1.6 2.6 0.10 0.09 0.05 23 .77 0 0.02 BBS EPD 104 29.6 0.50 1.8 2.4 0.13 0.09 0.28 25 .84 0 0.02 BBS EPD135CB 4.8 0.38 1.5 2.2 0.15 0.15 0.00 19 3.96 0 0.03 LATR
EPD200 29.4 0.57 2.1 2.6 0.18 0.16 0.10 30 1.02 0 0.03 LATR EPE18 30.0 0.54 2.5 3.7 0.18 0.18 0.00 33 1.10 2 0.02 PY EPE24 32.6 0.50 2.2 3.0 0.15 0.14 0.03 31 .95 0 0.02 PJ EPE 100 30.0 0.50 2.1 2.9 0.15 0.12 0.23 29 .97 0 0.02 BBS EPF47 29.6 0.64 2.4 3.1 0.19 0.18 0.06 33 1.12 0 0.02 PJ EPF49 30.0 0.57 2.4 3.3 0.15 0.14 0.05 34 1.13 1 0.03 PY EPFBCQ 30.0 0.50 2.2 2.6 0.14 0.1 1 0.26 31 1.03 2 0.02 PJ EPFl 00 29.5 0.71 2.2 2.7 0.18 0.19 -0.07 31 1.05 0 0.02 PJ RCC8 29.0 0.64 2.4 3.1 0.23 0.20 0.11 33 1.14 0 0.02 PJ RCDlO 28.5 0.71 2.4 2.7 0.23 0.22 0.08 34 1.19 0 0.02 PJ RCD121 29.9 0.50 2.1 3.0 0.17 0.15 0.11 29 .97 0 0.03 BBS RCE25 29.3 0.64 2.1 2.6 0.19 0.19 -0.01 30 1.02 0 0.02 PJ EPBPRR 29.5 0.64 2.1 2.7 0.20 0.15 0.25 30 1.02 1 0.03 PJUCA Mean 29.3 0.53 2.2 2.9 0.17 0.16 0.08 29.9 1.1 1 0.29 0.03 SD 5.99 0.12 0.3 0.3 0.03 0.03 0.02 4.35 0.55 0.64 0.01
(F,,), theta-p, and theta-s (two measures of F,,) following methods of Weir had the five highest values of standardized TA, P, H,, or mean genetic distance (1996) as implemented by Lewis and Zaykin (2000).The statisticf, the coan- to other populations, or the five lowest values for f. Tied values resulted in cestry coefficient or fixation index, represents departures from Hardy-Wein- selection of more than five sites for some measures. Populations ranking fa- berg equilibrium expectations within individual sites; F represents such de- vorably in multiple categories were considered to be a higher priority for viations over all sites. Theta-p represents the proportion of genetic differen- conservation of genetic diversity. tiation among populations from different vegetation types and was used to examine the relationship between genetic diversity and community compo-
sition. Theta-s represents the proportion of total genetic variation partitioned among sites in relation to the total variation present, treating both loci and
Erigeron parishii holds a considerable amount of genetic
sites as samples. We calculated 95% confidence intervals for these estimates
variation at the allozyme loci surveyed. At the species level,
from 5000 bootstrap replicates across loci. Gene flow, the number of migrants
nine of the 14 loci examined were polymorphic (P = 0.64).
per generation (N,,),was estimated both as N,,= 0.25(1 -theta-s)/theta-s
A total of 61 alleles were detected for all loci with an average
(Weir, 1996) and by using Slatkin's (1985) private allele method.
of 4.3 alleles per locus (SD = 1.94) and 5.2 alleles per poly-
Nei's (1978) unbiased genetic distance and geographic distance were com-
morphic locus (SD = 1.39). Species-level He was 0.19 (SD =
puted for all pairwise combinations of sites. The average genetic distance
0.17) and H,,was 0.15 (SD = 0.1 3).
from each site to all other sites was also calculated to determine whether any
The proportion of loci polymorphic within populations
sites were particularly unique. The strength of the relationship between the ranged from 0.29 to 0.71 with a mean of 0.53 (SD = 0.12)
genetic and geographic distance matrices was evaluated with a standardized Mantel statistic (Sokal and Rohlf, 1995) using PC-ORD (McCune and Mef- (Table I). The total number of alleles per population averaged ford, 1999). The significance of the Mantel statistic was assessed through a 29.9 (SD = 4.4) and ranged from 19 (EPD135CB, sampled randomization test using 1000 Monte Carlo simulations. from a small, peripheral population) to 38 (EPB203, from a
Conservation significance of individual sites based on diversity and differ- large, centrally located population). The second smallest num- entiation was assessed by identifying sites that supported private alleles and ber of alleles (20) was found at site EPB54, another peripheral
TABLE 2. Distribution of private alleles among 14 loci in 31 popula- tions of Erigemn l~iirishii.
Locu\ Allele Frequency Populailon
ACO l EPF49
XC03 EPD 12
PGDl EPE 18
PGM EPE 18
TPll EPC 13
TPI 1 EPC 13
population. When TA was standardized by sample size, the largest number of alleles was found in EPD135CB (3.96) and the smallest number was in EPB54 (0.67) (Table 1). The num- ber of alleles per locus within populations averaged 2.2 (SD = 0.3), and the number of alleles per poly~norphic locus av- eraged 2.9 (SD = 0.3). He within populations ranged from
0.10 to 0.23 and averaged 0.17 (SD = 0.03) (Table I). H, within populations was comparable in magnitude ranging from
0.09 to 0.22 and averaging 0.16 (SD = 0.03) over all popu- lations (Table I). Coancestry coefficients averaged over loci were positive within 26 of the 31 populations, indicating slight-to-moderate nonrandom mating (Table 1). Only four populations (EPE100, EPFBCQ, EPD104, and EPBPRR) ex- hibited substantial inbreeding (i.e., f > 0.20) and 71% of all populations had coancestry coefficients <0.1.
A total of ten private alleles was found in seven populations (Table 2). Three populations (EPCl3, EPEI 8, and EPFBCQ) had two private alleles each (Table 2). When found, these al- leles were always at low frequencies; in fact, all but two of these alleles occurred at frequencies c0.07. The remaining private alleles occurred at frequencies of 0.12 and 0.13. The distribution of private alleles did not have an obvious geo- graphic pattern (Table 2. Fig. 1).
Levels of genetic diversity were typically below average in the sample from site EPD135CB, the smallest population and smallest sample (Table I), and this site had the smallest num- ber of alleles. However, individuals at this site were not inbred (f = 0.0). While values of P, A, A,, and TA were below av- erage, when compared with equal sample sizes from other sites, levels of genetic diversity in EPD135CB were not dif- ferent from those sites. When N = 5 in all samples, the av- erage value of TA was 21.4 (SD = 2.1), P mias 0.41 (SD = 0.09), A was 1.5 (SD = 0.15), A, was 2.9 (SD = 0.24), He was 0.1 6 (SD = 0.04), H, was 0.14 (SD = 0.04). and f was
0.09 (SD = 0.17). Thus the apparently low levels of diversity at site EPD135CB (Table 1) appear to be the result of a small sample size, not the result of an especially genetically depau- perate, small population. In fact, this site had the largest stan- dardized TA. Additionally. it should be noted that no inbreed- ing was detected at this site. In fact, it ranked among the least inbred sites.
On average, 90% of the multilocus genotypes within sample sites were unique (SD = 7.3). This percentage ranged between 77% (EPDS and EPDI 04) and 100% (EPB26. EPB.55. EPDI 2, EPE100, RCC8, and RCDIO). The nonuniclue genotypes were not, howevel; all identical to one another. A total of 60 indi- viduals out of the 868 analyzed that had nonunique genotypes
TABLE3. F statistics for 12 loci (FESTI and FEST2 were cxcluded because they were monomorphic) in 31 populations of Erigeror~ pai.i.sl~iihierarchically arranged in vegetation types calculated using methods of Weir (1996). Theta-p represents differentiation of pop- ulations within vegetation types; theta-s represents differentiation among all populations. Standard deviations for each locus are in parentheses and were calculated by jackknifing over populations. Upper and lower bounds were calci~lated by bootstrapping across loci.
Lozu f F Theta-p 'l'hcia-c
ACO 1 AC02 AC03 SKDH PGD1 PGD3 L,AP PGM AATl TPI l TP12 UDP Overall Upper bound Lower bound
within their respective sites. These nonunique genotypes were never widespread within sites. For example, 80% of the non- unique genotypes within sites were each shared by only two individuals and 17% were shared by three individuals. At most, four individuals within sample sites shared a given ge- notype, and this only occurred twice. Further. it is likely that the proportion of unique genotypes was underestimated be- cause many individuals that were indistinguishable were miss- ing data at one or more of the 12 polymorphic loci.
The coancestry coefficient, ,f; averaged 0.08 across all loci and sites; based on confidence intervals derived from bootstrap estimates, f was significantly different from zero (Table 3). Coancestry coefficients for individual loci varied from f = -0.02 (AC02. TPI2, and UDP2) to ,f = 0.17 (LAP) (Table 3). These results indicate a slight heterozygote deficit in in- dividual sites due to nonrandom mating. Indicating deviations in heterozygosity in individuals relative to the all sites com- bined, F was also significantly greater than zero (Table 3), This deviation has contributions from both drift and nonrandom mating (Weir, 1996). Overall, population differentiation esti- mated from samples was moderate and significantly different from zero, with theta-s averaging 0.1 2 over all loci among all populations (Table 3). In other words, -88% of the diversity in this species was common to all populations. Based on this level of differentiation, the number of migrants among popu- lations per generation was estimated to be 1.8. The private allele method yielded a similar result, indicating 1.6 migrants per generation. In contrast, there uras no differentiation among populations in different vegetation types as theta-p averaged
0.003 and did not differ significantly from zero (Table 3).
Nei's unbiased genetic distance among all pairs of sampling sites averaged 0.028 (SD = 0.023) and ranged from 0.001 between EPC104 and EPB203 to 0.133 between RCD121 and EPD202, indicating little overall differentiation among sam- pling sites. Further, there was not much variation in the av- erage distance of one site to all other sites (Table 1). Geographic distance between all pairs of sample sites averaged
May 20011 NEEL AND ELLSTRAND-GENETICDIVERSITY IN ERIGERON 81 5
10.0 km (SD = 9.65) and ranged from 0.4 to 49.2 km. When the disjunct site EPBPRR was not included, the greatest dis- tance between sites was 27.3 km and the average distance was
7.0 kni (SD = 5.65). At the same time, the average genetic distance among populations did not change when EPBPRR was excluded (0.028 [SD = 0.021). The Mantel test indicated that there was no significant correlation between geographic distance and genetic distance among all pairs of sample sites (r2= 0.01, P = 0.32). Thus, there was no apparent geographic pattern to genetic distance among populations.
Twenty-two sites ranked at least once in the top five values for five measures of diversity and one measure of differenti- ation or the bottom five values for inbreeding or had private alleles (Table 1). Fourteen sites ranked favorably more than once and six sites ranked riiore than twice (Table I), although the measures of diversity with high ranks in these five sites varied. Sites EPD12 (4), EPB26 (3), EPB31 (3), EPD135CB (3), RCDlO (3), and EPBPRR (3) all ranked favorably more than twice.
Given its endemism and history of fragmentation, E. pnri- shii supports a surprisingly large amount of genetic variation. Values for all genetic diversity rneasures were higher than those typical for narrowly endemic plant taxa (Hamsick and Godt, 1989; Loveless and Hamrick, 1989; Hamsick et al., 1991;Godt, Johnson, and Hamsick, 1996). In fact, all species- level measures except He were higher than mean values for 329 dicotyledonous species (reviewed by Hamsick and Godt 1989). Species-level He for E. parisltii was between means reported by Hamrick and Godt (1989) in 81 endemic (He = 0.096) and 101 narrowly distributed (He = 0.137) taxa. Most of the observed diversity was held within populations and, as such. within-population diversity estimates were also higher than those expected for narrowly endemic plant taxa. or example, P, A, and A, within populations of E. pnrislzii were higher than the average values for dicots as well as widespread taxa in general (Hamrick and Godt, 1989). Within E. pari,rhii populations, He was more similar to levels found in regionally distributed taxa. One possible explanation for these discrep- ancies in expected letels of divkrsity is that, compared to many endemic plants, E. pnrishii populations can be large and extensive. Another explanation is that the species is not-yet in evolutionary equilibrium and that the high levels of variation reflect a recent time before fragmentation when populations were less isolated.
The smallest population (EPD135CB) did have the smallest number of alleles and below average proportion of polymor- phic loci, however, measures of heterozygosity and inbreeding were not significantly lower than average measures from all populations. Furthermore, when compared with the same sam- ple sizes from larger populations, the total number of alleles in EPD1325CB was not particularly low. Thus, while this pop- ulation is genetically depauperate, it is most likely due to ef- fects of sample size rather than inbreeding or drift. This result is curious given the large amount of theoretical and empirical evidence predicting severe effects of inbreeding and drift in small populations (reviewed in Barrett and Kohn, 1991; Ells- trand and Elam, 1993). The amount of genetic diversity and lack of inbreeding observed indicate that this population may represent a recent colonization event. The seven individuals at this site were found in an ephemeral wash at the base of Black- hawk Mountain near the lower elevational limit for the taxon. Erigemn parishii had not been documented from this site prior to 1995 when five individuals were found in a vegetation sam- pling plot. A number of populations occur upslope from this location and seeds could have washed downstream during a rainstorm. Erigeron pnrishii typically occupies washes and is thought to potentially disperse in this manner, although no di- rect evidence regarding dispersal mechanisms exists. The clos- est potential source population is -0.8 km from and at an elevation 60-80 m above site EPD135CB. It is possible that individuals do not persist at this site but rather the site is col- onized from larger populations during favorable years. It would be worthwhile to follow this population over time to see how long it persists and to determine the frequency and source of newly recruited individuals.
We found no evidence for extensive agamospermy or clonal reproduction in E. pnrishii. If these reproductive modes were common, we would expect a large proportion of individuals within samples to share n~ultilocus genotypes (Ellstrand and Roose, 1987). In fact. on average, at least 90% of the multi- locus genotypes in each sample were unique (at least 77% of genotypes within sites were unique). Most (80%) of the non- unique genotypes were shared by only two individuals. In comparison, Noyes and Soltis (1996) found that only 10% of individuals in seven agamospermous populations of Erigeron cor~zposit~~s
had distinguishable genotypes. In the same study, an average of 60% of the individuals in sexually reproducing populations had unique genotypes. They also found that pop- ulations were highly differentiated from one another, which also strongly contrasted with our results. Based on the high diversity as well as low levels of inbreeding and population differentiation we conclude that E. pnrishii likely primarily reproduces sexually through outcrossed matings (Heywood, 1991).
The levels and patterns of diversity documented also allow us to make inferences about some aspects of the demographic history of this taxon (Milligan, Leebens-Mack. and Strand, 1994). The large percentage of polymorphic loci and numbers of alleles oer locus indicate that E. r~arishiihas no historv of sufficiently severe or long-lasting population bottlenecks to cause loss of genetic diversity. Large historical populations are also indicated by the observed low to moderate levels of in- breeding (Table 1). The lack of major differentiation among populations (Table 3) suggests either substantial gene flow among populations or that fragmented populations have not yet been isolated long enough for genetic drift (or selection) to have caused population differentiation (Ellstrand, 1992).
Because most populations of E. par-ishii have likely remained large throughout most of their history or have been interconnected at least until the recent past, severely reducing population si~es could have detrimental effects. Such reduc- tions could alter the mating structure within populations by increasing selling or mating among relatives and thus increase total inbreeding and potentially inbreeding depression (Hus- band and Schemske, 1996). Fragmentation isolating previously connected populations could also have deleterious effects by disrupting gene flow among populations and could exacerbate loss of diversity through drift and further increase inbreeding in the remaining fragments (Templeton et al., 1990; Young. Boyle, and Brown, 1996). Conservation efforts focusing on maintaining large populations of E. pnrishii as well as con- nections among those populations should be sufficient to sus- tain the existing high levels of diversity in this taxon and to
816 AMERICANJOURNALOF BOTANY [Vol. 88
prevent the deleterious effects described above. In addition to protecting existing sites, restoration of degraded sites to in- crease population sizes and to connect fragmented populations could be useful but should be a lower priority than protecting existing sites. However, there is no evidence that reintroduc- tion is necessary to "genetically rescue" or supplement any population, a measure that was recommended in the draft re- covery plan for this taxon (USFWS, 1997).
Beyond general guidelines, patterns of genetic diversity re- vealed by molecular markers can contribute to setting conser- vation priorities among populations. While there is general agreement that conserving genetic diversity is important, there is little agreement on the specific measure of diversity that should be targeted. Heterozygosity and inbreeding are typi- cally targeted because of the direct influence of levels of di- versity on fitness (reviewed in Huenneke, 1991; Linhart and Grant, 1996; Latta and Mitton, 1997). Evidence for short-term fitness benefits of heterozygosity is accumulating (Frankel and SoulC, 1981 ; Ellstrand and Elam, 1993; Oostermeijer, Van Eijck, and Den Nijs, 1994; Oostermeijer, Van Leeuwen, and Den Nijs, 1995) but is not ubiquitous (e.g., Ritland, 1990; Eguiarte, Perez-Nasser, and Pifiero, 1992; Ouborg and van Treuren, 1995; Lynch, 1996). Effects of heterozygosity on fit- ness are most apparent when comparing selfed and outcrossed offspring (e.g., Waller, 1984; Holtsford and Ellstrand, 1990) or when sampling across age classes within populations (e.g., Schaal and Levin, 1976; Murawski, Nimal Gunatilleke, and Bawa, 1994; Allphin, Windam, and Harper, 1998).
Recently, Petit, Mousadik, and Pons (1998) strongly advo- cated using allelic richness (standardized for different sample sizes using rarefaction) as allelic diversity defines the ultimate limits of evolutionary potential. Maintaining the evolutionary potential of a species is thought to increase the probability of persistence of a taxon (Storfer, 1996). In addition, Petit, El Mousadik, and Pons (1998) suggested evaluating the signifi- cance of individual populations based on both levels of allelic richness and distinctiveness. While conserving markers is not the ultimate goal, they do provide an easily measured index of general levels of diversity. However, because patterns of allozyme diversity are not always concordant with other types of diversity they thus must be applied cautiously (Hamrick, 1989; Lynch, 1996; Storfer, 1996).
Because all populations in this study were relatively diverse and there was relatively little differentiation among sampling sites (Table I), individual measures of allozyme diversity pro- vided limited means of prioritizing any particular populations for conservation. For example, 71% of the samples ranked favorably in at least one measure of diversity, distinctiveness, or inbreeding. To distinguish among sites we considered those that ranked highly in more than two measures to be of partic- ular conservation priority. These sites included EPD12 (4), EPB26 (3), EPB31 (3), EPD135CB (3), RCDlO (3), and EPBPRR (3). Interestingly, only two of these samples (EPDI 2 and EPBPRR) had any private alleles, another measure of the contribution of a population to diversity within a taxon. Fur- thermore, some populations high in some genetic diversity measures were low in others. For example, EPBPRR ranked highly in terms of P, He, and genetic distance and supported one private allele, but was one of the most inbred of all sites (Table 1). Clearly, the rank of a given population depends on the measure of genetic diversity examined. We would not ad- vocate conserving only the six populations that ranked highly in multiple measures of diversity as they would not represent the range of E. parishii and thus would not likely include the range of adaptive variation in the taxon. Rather, we recom- mend including populations represented by these samples in a reserve network that included populations from throughout the ecological and geographic range of the taxon. Inclusion of particular populations in a reserve network should also take into account population size and extent, and habitat quality and defensibilitv.
In summary, E. parishii currently supports a substantial amount of allozyme variation at both the species and popu- lation levels. Additionally, rates of inbreeding and drift appear to be moderate to low. Because loss of genetic diversity is not of immediate concern for E. parishii and because populations were not highly differentiated, reserve designs based on eco- logical factors would likely suffice to maintain genetic diver- sity. For example, protecting sites throughout the ecological and geographic range of E. parishii that incorporate the allo- zyme diversity documented would probably encompass its range of adaptive variation, although that variation was not characterized here. The high levels of diversity combined with the low levels of differentiation among populations support maintaining a network of large, interconnected populations. Such a network would prevent changes in mating system struc- ture that would increase selfing and biparental inbreeding and increase rates of drift. Given the low levels of differentiation, it would not be necessary to conserve a large number of pop- ulations simply to represent the genetic diversity in terms of including all alleles sampled. Howevel; more numerous, large, interconnected populations would be necessary to maintain the processes that support that diversity. This could be accomplished through a combination of protecting existing sites and restoring degraded habitat. The data presented here provide a valuable baseline for future comparisons of genetic diversity to evaluate effectiveness of protected areas and restoration in maintaining genetic diversity or evaluating consequences of further fragmentation and population loss.
ALLPHIN,L., M. D. WINDHAM, AND K. T. HARPER. 1998. Genetic diversity and gene flow in the endangered dwarf bear poppy, Arcrornecon humilis (Papaveraceae). A~nericnn Jo~ournnl of Botany 85: 125 1-1261.
BARRETT, S. C. H., AND J. R. KOHN. 1991. Genetic and evolutionary con- sequences of small population size in plants: implications for conserva- tion. In D. A. Falk and K. E. Holsinger [eds.], Genetics and conservation of rare plants, 3-30. Oxford University Press. New York, New York, USA.
BROWN, A. H. D. 1989. Genetic characterization of plant mating systems. In A. H. D. Brown, M. T. Clegg, A. L. Kahler, and B. S. Weir [eds.], Plant population genetics, breeding, and genetic resources, 145-162. Sin- auer, Sunderland, Massachusetts, USA.
CALZIA,J. P. 1993. Mineral resource assessment of the San Bernardino Na- tional Forest, California; executive summary and a non-technical summary for land and resource managers. United States Geological Survey Open File Report 93-0552.
CESKA,J. E, J. M. AFFOLTER, AND J. C. HAMRICK. 1997. Developing a sampling strategy for Bnprisin nrzlclzrzfem based on allozyme diversity. Consei~~ntiorz
Biolog)~11: 1133-1 139.
CLEGG, M. T., ET AL. 1995. Science and the Endangered Species Act. Com- mittee on Scientific Issues in the Endangered Species Act. National Re- search Council. National Academy Press, Washington, D.C., USA.
COOPERRIDER, 1991. Conservation of biodiversity of western rangelands.
A. Iiz W. E. Hudson [ed.], Landscape linkages and biodiversity, 40-53. Island Press, Washington, D.C., USA.
EGUIARTE, AND D. PINERO. 1992. Genetic structure.
L. E.. N. PEREZ-NASSER, outcrossing rate and heterosis in Astrocaiyzrrn ilzenicnrzllnz (tropical
May 20011 NELL AND ELLSTRAND-GENETICI-'ARISHII 817
DIVERSITY IN ERIG~RON
palm): ilnplicatio~is for evolution and conservation. Heredip 69: 217-
228. EL.LSTRAND,N.C. 1992. Gene flow among seed plant populations. .NCMJ Forests 6: 241-256.
-, AND D. R. ELAM. 1993. Population genetic consequences of small population size: implications for plant conservation. Aiz~lrnl Re~biei.~ of Ecology ai~d Systemntics 24: 217-242.
-----, AND M. L. ROOSE. 1987. Patterns of genotypic diversity in clonal plant species. Ar?zericnn Jo~~nzal of Hotairy 74: 123-1 11.
FALK, D. A. 1992. From conservation biology to conservatio~l practice: strat- egies for protectiug plant diversity. liz P. L. Fiedler and S. K. Jain [eds.], Conservation biology: theory and practice of nature conservation and management. 396-43 I. Chapman and Hall, New York, New York, USA.
FRANKEL:0. H., AND M. E. SOULE. 1981. Conservation and evolution. Cam- bridge University, Cambridge, UK. GAINLS, M. S., J. E. DIFFENDORFER. ANII T. S. WHITTAM.
R. H. TAMARIN, 1997. The effects of habitat fragmentation on tlie genetic structure of srilall mammal populations. Jozoalrr~alqf He~edity88: 294-304.
Goor, M. J. W., B. R. JOHNSON. AND J. L. HARIRICK. 1996. Genetic diversity and population size in four rare southern Appalachian plant species. Conservation Biology 10: 796-805.
HAMRICK,J. L. 1989. Isozymes and the analysis of genetic structure in plarit populations. Irr D. E. Soltis and I? S. Soltis [eds.], Isozyrnes in plant biology, 87-105. Dioscorides, Portland, Oregon, USA.
----, .AND M. J. W. GODT. 1989. Allozyme diversity in plant species. lii
A. H. D. Brown, M. T. Clegg, 4. L. Kahler, and B. S. Weir [eds.]. Plant population genetics. breeding and genetic resources, 43-63. Sinauer, Sunderland, Massachusetts, USA.
--. D. A. MURAWSKI, AND M. D. LOVELESS. 1991. Correla- tions between species traits and allozyme diversity: implications for con- servation biology. Ii7 D. A. Falk and K. E. Holsinger [eds.], Genetics and conservation of rare plants. 75-86. Oxford University, Nev York, Ne\% York, USA.
HARRISON. AND J. MCNEELY. 1984. The worid coverage of J., K. MILLER, protected areas: development goals and enviroumental needs. In J. A. McNeely and K. R. Miller [eds.], National parks, conservation and de- velopment: the role of protected areas in sustaining society, 24-33. Smithsonian Institution, Washington, D.C., USA. HEYWOOD,J. S. 1980. Genetic correlates of edaphic differentiation and en- demism in Gaillardin. Ph.D. dissertation, University of Texas, Austin, Texas, USA. 1991. Spatial analysis of genetic variation in plant populations. Arz111rn1Reviei.1~ of Ecology cirzd Systenlotics 22: 335-355. HIC~IAN. 1993. The Jepson manual. Uiliversity of California, Berkeley,
HOLSINCER,K. E., AND L. D. GOTTLIEB. 1991. Conservation of rare and endangered plants: principles ant1 prospects. Irz D. A. Falk and K. E. Holsinger [eds.], Genetics and co~iservation of rare plants, 149-170. Ox- ford University Press, New York, New York, USA.
HOLI.SFORD,T. P., AND N. C. ELLSTRAND. 1990. Inbreeding effects in Clarkin teir~hlorieiisis(Oiiagraceae) populations with different natural outcrossing rates. Evolution 44: 203 1-2046.
HUENNEKE, 1991. Ecological implicatio~ls of genetic variation in plant
L. E populations. Irr D. A. Falk and K. E. Holsinger [eds.], Genetics aud conservation of rare plants, 31-44. Oxford University Press, New York, New York. USA.
HUSBAND,B. C.. AKD D.W. SCHERISKL. 1996. Evolutiou of the magnitude and tinling of illbreeding depression in plants. Evolutioir 50: 54-70.
LATI-A,R. G., AND J. B. MITTON. 1997. A comparison of population differ- entiation across four classes of gene markers in limber pine (Piizttsjesilis James). Genetics 146: 11 53-1 163.
LEWIS,P. O., AND D. ZAYICIN. 2000. Genetic data analysis: computer pro- gram for the analysis of allelic data. version 1.0 (d15). Free program distributed over the internet from the GDA Home Page at http://alleyn. eeb.uconn.edu/gcia/.
Y. B., AND M. C. GRANT. 1996. Evolutionary sig~lificauce of local genetic differentiati011 in plants. Airrlltal Review of Ecologj arld Systeirz- atics 27: 237-277.
LOVELESS,M. D., AND J. L. HARIRICK. 1989. Ecological determinants of genetic structure in plant populations. ili~irl~nl Ecologj aii~l
Review of Svsten7atics 15: 65-95. LYNCII,M. 1996. A quantitative-genetic perspecti\-e on conservation issues. In J. C. Avise and J. L. Hamrick [eds.], Co~lservation genetics: case
histories from nature, 471-482. Chapman and Hall, New York, New
MCCLIYE.B., AND M. J. MEFFORD. 1999, PC-ORD. Multivariate analysis of ecological data, versiou 4. MJM Software Design, Gleneden Beach, Oregon, USA.
MILLIGAN, AND A. E. STRAND. 1994. Co~lservation
B. G., J. LEEBENS-MACK, genetics: beyond the maintenance of maker diversity. Moleculnr. Ecology
3: 423-435. MOLDEKKE,A. R. 1976. California polli~iation ecology and vegetation types.
Phytologia 34: 305-36 1. MURAWSKI, GUNATILLEKE,
D. A,, I. A. U. NI~IAL AND K.S. BAWA. 1994. The effects of selective logging on inbreeding in Shoreci nrc~gistophylln (Dipterocarpaceae) from Sri Lanka. Coizseri:arion Biolog), 8: 997-1 002.
NEEL,M. C. 2000. The structure of diversity: i~nplications for reserve design. Ph.D. dissertation, University of California, Riverside, California, US4. NEI, M. 1978. Estimation of average heterozygosity and genetic distance fro111 a small number of individuals. Gerretics 89: 583-590.
Noss, R. F., M. A. O'CONNELL, AKD D. MURPHY. 1997. The science of conservation planning: habitat couservation uuder the Endangered Spe- cies Act. Island Press, Covelo, California, USA.
NOYES, R. E., AND D.E. SOLTIS. 1996. Genotypic variatiou in agatnosper- mous Eiigeroil coi,r11ositrf.s (Asteraceae). Airrer.ici~n Jourrzcrl ($Botany 83: 1292-1 303.
O'MALLEY,D., N. C. WHEELER, AND R. I? GURIES. 1980. A manual for starch gel electrophoresis. Staff Paper Series 1 I. Department of Forestry, College of Agricultural and Life Sciences, University of Wisconsin, Madison, Wisconsin, USA.
J. G. B., M. W. VAN EIJCK, AVD J. C. M. DEK NIJS. 1994. Analysis of the relationship between allo~y~ne
heterozygosity and fitness in the rare Gerztiaizn ,~~rie~iiizorrantlie
L. .lozt~rral of Evolr~tioiraiy Biologj
8: 739-759. -. N. C. VAN LEELT\VLK,
AKD J. C. M. DEN NIJS. 1995. Offspring fitneqs in relation to population size and genetic variation in the rare perennial plant species GentWrza prle~tnrotrnntlre L. Oecologin 97: 289-
296. OPDARI,P., R. FOPPEN, R. REIJNAN, AND A. SCHOT~IAN.
1994. The landscape ecological approach in bird conservation: integrating the metapopulation concept into spatial planning. Ibis 137: S 139-S146.
OCBORG, N. J.. AND R. VAK TREUREN. 1995. Variation in fitness-related characters among small and large populatiolls of Snliia ptziterzsis. Jozour- ~ral of Ecology 83: 369-380.
PETIT, R. J., A. EL MOUSAIIIK, AND 0. PONS. 1998. Identifying populations for conservation on the basis of genetic markers. Corlsei.~'ntion Biologj
12: 844-855. RICHARDS,A. 1999. Plant breeding systems. 2nd ed. Chapman and Hall, London, UK. RITLAND,K. 1990. 111fereuces about inbreeding depression based on chatiges of the illbreeding coefficient. Evolution 44: 1230-1241. SCHAAL,B. A,, AND D. A. LEVIS. 1976. The demographic genetics of Licirris
cyliildracen Michx. Anzericnrr ivaruralist 1 10: 191-206. SCHEAISKL, B. C. HUSBAND: M. H. RIJCKELSHAUS.
D. W.. C. GOODWILLIE,
I. M. PARKER. AND J. G. BISHOP. 1994. Evaluating approaches to the conservation of rare and endangered plants. Ecology 75: 584-
606. SIMBERLOPF, 1988. The co~itribution of population and community biol-
D. ogy to conservation science. Arznztul Reilie~v of Ecology oilil Systenzntics
19: 473-5 11.
SKINNER,M. U',,AND B. hI.PAVLIK. 1994. Inventory of rare and endangered vasciilar plants of California. Califor~iia Native Plant Society. Sacramen- to, California, USA.
SLATKIN,M. 1985. Rare alleles as indicators of gene flow. Evolirrioi~39: 53-
65. SOKAL, R. R., AND E J. ROIILF. 1995. Biometry, 3rd ed. W.H. Freeman, San Francisco, California, USA.
STORFER,A. 1996. Qualititalive genetics: a promiqing approach for tlie as- sessment of genetic variation in endangered species. TWILLIS
in Ecology aizrl Evolirtiorl l l : 343-348.
TEMPLETON,A. R., K. SHAW, E. ROUTMAN, AND S. K. DAVIS. 1990. The genetic consequences of habitat fragmentation. Airrrnls of tire h1issolir.i Borcinic Gnrclerr 77: 13-27.
T~on?~soN, 1996. Evolutionary ecolog!, atid the conservatio~i of bio- J. N.
diversity. Treizcls iri Ecolog~l rirzd E~~olzrtiorr l I: 300-303. USFWS [U.S. FISH AND WILDLIFESER~ICE].1994. Final Rule. Elidangered and threatened wildlife and plants: five plants from the San Berliardino Mountains in southem California determined to be threatened or endan- gered. 50 CFR Part 17. Fedeml Register. 59: 43 652-43 664.
1997. San Bernardino Mountains carbonate endemic plants draft recovery plan. U.S. Fish and Wildlife Service Region 1, Portland, Oregon, USA.
U'ALLER,D. M. 1984. Differences in fitness between seedlings derived from cleistogamous and chasmogamous flowers in I~~zpatieiisciipensis. Evo- lution 38: 427-440.
WEEDER',N. E, AND J. E: WEKDEL. 1989. Visualization and interpretation of plant isozymes. Iiz D. E. Soltis and E! S. Soitis [eds.], Isozymes in plant biology, 5-45. Dioscorides Press, Portland, Oregon, USA.
WEIR,B. S. 1996. Genetic data analysis 11. Sinaues, Sunderland, Massachu- setts, USA.
WILCOVE, D. S., D. ROTHSTEIN,J. DUBOW, A. PHILLIPS, AR'D E. LOSOS. 1998. Quantifying threat5 to imperiled species in the United States. BioScierlce 48: 607-6 15.
YOUNG, A,, T. BOYLE, AND T. BROWN. 1996. The population genetic con- sequences of habitat Gagmentation for plants. Trerzils irl Ecology ant1 Evol~/ti(~il