Embed Size (px)
Citation preview
BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofitpublishers, academic institutions, research libraries, and research funders in the common goal of maximizing access tocritical research.
Mating System Analysis of Alnus maritima (Seaside Alder), aRare Riparian TreeAuthor(s): James M. Jones and J. Phil GibsonSource: Castanea, 77(1):11-20. 2012.Published By: Southern Appalachian Botanical SocietyDOI: http://dx.doi.org/10.2179/11-024URL: http://www.bioone.org/doi/full/10.2179/11-024
BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in thebiological, ecological, and environmental sciences. BioOne provides a sustainable onlineplatform for over 170 journals and books published by nonprofit societies, associations,museums, institutions, and presses.
Your use of this PDF, the BioOne Web site, and all posted and associated contentindicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use.
Usage of BioOne content is strictly limited to personal, educational, and non-commercialuse. Commercial inquiries or rights and permissions requests should be directed to theindividual publisher as copyright holder.
Mating System Analysis of Alnus maritima(Seaside Alder), a Rare Riparian Tree
James M. Jones1 and J. Phil Gibson1,2*1Department of Botany-Microbiology, University of Oklahoma, 770 Van Vleet Oval,
Norman, Oklahoma 730192Department of Zoology, University of Oklahoma, 730 Van Vleet Oval,
Norman, Oklahoma 73019
ABSTRACT Effective conservation requires an understanding of the genetic interactionsamong populations and individuals of a species, particularly those with fragmented, isolateddistributions. Alnus maritima (seaside alder) is a rare tree species with an extremely fragmenteddistribution of highly isolated populations in the Delmarva Peninsula, Georgia, and Oklahoma.We conducted a mating system study to estimate the outcrossing rate, inbreeding coefficient,biparental inbreeding rate, and correlation of paternity in progeny from a Georgia and anOklahoma population to investigate the effects of isolation on the A. maritima mating system.Data from nine microsatellite loci showed similarly high multilocus outcrossing rates in bothpopulations (tm 5 0.94). Individual tree outcrossing rates were also high (tm 5 0.873–1.047). Therewas no significant biparental inbreeding in either population, but there was significantly highercorrelated paternity in the Oklahoma population. Results showed the high outcrossing expectedfor a wind-pollinated, monoecious species that can promote the maintenance of genetic variationdetected in A. maritima seed pools and standing populations. Likewise, pollen flow amongOklahoma populations may promote maintenance of regional genetic variation. However,despite the genetic diversity in the seed pool generated by A. maritima’s highly outcrossed matingsystem, failure of new individuals to be recruited into populations from seed presents an obstaclethat will need to be considered when developing conservation strategies for this rare species.
Key words: Alnus, fragmented populations, Georgia, mating system, Oklahoma, paternityanalysis.
INTRODUCTION The mating system andability for populations to genetically interactvia gene flow are critical factors that must beconsidered in conservation and managementplanning for plant species with fragmented,isolated populations. Fragmentation is predict-ed to reduce gene flow, which can lead to loss ofgenetic variation, increased inbreeding andinbreeding depression, and significant geneticdifferentiation among populations as popula-tions become smaller, more fragmented, andmore isolated (Loveless and Hamrick 1984,Hamrick and Godt 1990). Consistent withmany of these predictions, studies of plantspecies with natural or anthropogenically frag-mented distributions have typically shown the
expected genetic differentiation among frag-ments (Gibson and Hamrick 1991, Gibson andWheelwright 1995, Cardoso et al. 1998, Jonesand Gibson 2011), reduced gene flow (Sork andSmouse 2006, Bittencourt and Sebbenn 2007),increased susceptibility to genetic drift (Barrettand Kohn 1991, Karron 1991), and higher levelsof inbreeding (Barrett and Kohn 1991, Ellstrandand Elam 1993). These population geneticconsequences of fragmentation are particularlyimportant for species that are targets forconservation because the loss of genetic diver-sity and increase in frequency of deleteriousalleles can potentially limit the ability ofpopulation fragments to respond and adapt tochanging environments or shift their range inresponse to global climate change (Davis andShaw 2001, Jump and Penuelas 2006, Sork andSmouse 2006, Coates et al. 2007, Kelly andGoulden 2008, Thuiller et al. 2008).
*email address: [email protected]
Received July 6, 2011; Accepted September 2, 2011.
DOI: 10.2179/11-024
CASTANEA 77(1): 11–20. MARCH 2012Copyright 2012 Southern Appalachian Botanical Society
11
The present study investigates the matingsystem of Alnus maritima (Marsh.) Muhl. exNutt (seaside alder), a rare, woody, perennialspecies that has the most highly disjunctdistribution of any tree species in NorthAmerican (Little 1975, Stibolt 1981). Seasidealder grows in three widely separated loca-tions: the Delmarva Peninsula region ofDelaware, Maryland, and Virginia; south-central Oklahoma; and northwest Georgia(Figure 1). We conducted a mating systemanalysis of seaside alder to estimate the levelsof outcrossing, biparental inbreeding, andcorrelated paternity in a Georgia and anOklahoma subspecies. These populationswere chosen because they are of similar sizeand are clearly delimited as distinct popula-tions, thereby providing an opportunity tocompare the mating system among trees that(a) are in populations with differing degreesof isolation, and (b) grow in areas with
differences in vegetation that may influencepollen dispersal by wind. We expect popula-tion and family-level outcrossing rates to behigh in both study populations due to seasidealder’s monoecious gender and anemophi-lous pollination system, but that biparentalinbreeding and correlated paternity will begreater in sites with dense vegetation thatmay limit pollen movement. Results of thisstudy will provide fundamental informationon seaside alder’s mating system that canguide conservation planning for this species.
MATERIALS AND METHODSStudy Species Distribution and Reproductive
PhenologyRegional seaside alder populations have beendesignated as three separated subspecies(Schrader and Graves 2002). In Delmarva,Alnus maritima subsp. maritima grows alongthe Nanticoke, Wicomico, Pocomoke, and
Figure 1. Distribution of Alnus maritima subspecies in Delaware (DE), Maryland (MD), Georgia (GA), andOklahoma (OK).
12 CASTANEA VOL. 77
Choptank Rivers and their tributaries inMaryland and western Delaware. It alsooccurs sporadically in small clusters or isolat-ed trees around ponds of dammed creeks andstreams. In Oklahoma, Alnus maritima subsp.oklahomensis occurs along the Blue River,Delaware Creek, Pennington Creek, and othersmall tributaries in Johnston and PontotocCounties. A single population of Alnus mar-
itima subsp. georgiensis grows in DrummondSwamp, a spring-fed swamp near Eurharleein Bartow County, Georgia. Population ge-netic analyses of the Delmarva, Oklahoma,and Georgia seaside alder populations haveindicated that plants growing in these areasare members of relictual populations that arelikely the remnants of what was once a morewidespread distribution in North America(Gibson et al. 2008, Jones and Gibson 2011).
Due to its rarity and high degree ofisolation, the International Union for Conser-vation of Nature (IUCN) has given A. maritima
a G3 conservation, indicating the species isvulnerable, near threatened, and faces a highrisk of extinction throughout its range in themedium term (IUCN 2010, NatureServe2010). At the state level, the Delmarvasubspecies is given a conservation rank ofS3, indicating it is potentially vulnerable toextinction in Delaware and Maryland (Dela-ware Natural Heritage Program 2003, Mary-land Department of Natural Resources 2004).The Oklahoma subspecies has an S2 rank andis considered imperiled (Oklahoma BiologicalSurvey 2010). The Georgia subspecies iscritically imperiled and has the highest S1conservation ranking (Georgia Department ofNatural Resources 2006).
Alnus maritima trees have a multistemmedgrowth form with individuals growing asclearly delimited clumps. It is the only NorthAmerican species in the small subgenusClethropsis. Other North American alders arepredominantly in the subgenus Alnus. Like allalders, A. maritima is monoecious and pro-duces flowers in unisexual catkins. However,while other alders flower in the spring beforetrees produce leaves and release their seeds inautumn of the same year, A. maritima andother members of Clethropsis follow a pheno-logical cycle in which catkins open forpollination in late summer or early autumnwhile individual trees and the surrounding
vegetation are full of leaves. The seeds donot mature for release until autumn of thefollowing year. Most seeds disperse in late fallor early winter, but some can persist in femalecatkins for an additional year (Schrader andGraves 2000, Chen and Li 2004, Gibson et al.2008).
Sampling SitesSamples were collected at one study site inOklahoma and one in Georgia. In Oklahoma,samples were collected from the large popu-lation of approximately 250 trees growingalong the Blue River in the Blue River WildlifeManagement Area in Johnston County, Okla-homa (34u21935.420N, 84u54954.200W). Thispopulation is one of five known seaside alderpopulations occurring along different streamsin the region. Populations are separated fromone another by approximately 5–10 km witha heterogeneous landscape of agriculturalareas, native grasslands, and upland forestsincapable of supporting A. maritima betweenthem. Trees along the Blue River and otherareas in Oklahoma grow among dense ripar-ian vegetation of Populus deltoides Bartram exMarsh. Salix spp., Carya spp., and other bot-tomland tree species. In contrast, the Drum-mond Swamp population (34u7953.090N,84u54954.200W) is the only population inGeorgia and is isolated from the nearest A.
maritima population by more than 1,300 km.The Georgia population is composed of twodistinct clusters of individuals at opposite endsof Drummond Swamp. The clusters, hereafterreferred to as Drummond Swamp East andWest, contain approximately 70 and 200 trees,respectively. The clusters are separated byapproximately 0.5 km but with no landscapefeatures to physically obstruct pollen flowbetween them. Trees in Drummond SwampEast are scattered in open-water areas with nosurrounding vegetation except small clumps oflow aquatics (e.g., Saggitaria spp., Juncus spp.).In Drummond Swamp West, A. maritima treespredominantly grow in a monospecific standsurrounding the spring that is the primarywater source for Drummond Swamp, with a fewscattered trees in open areas. A majority of treesin both Georgia and Oklahoma sites weregenotyped as part of a previous populationgenetic study of A. maritima (Jones and Gibson2011).
2012 JONES, GIBSON: MATING SYSTEM ANALYSIS OF SEASIDE ALDER 13
Sample Collection and DNA ExtractionWe initially collected 5 to 10 mature catkinsfrom 10 trees in the Drummond Swampsubpopulation (five trees in DrummondSwamp East, five trees in Drummond SwampWest) and 8 trees in the Blue River popula-tion. Female catkins were returned to the labfor seed extraction. We removed seeds fromfemale catkins in the lab and pooled seedsfrom each maternal tree. We chose 30 filledseeds at random from each maternal tree forsubsequent DNA extraction. Successful DNAextraction from a minimum of 20 filled seedswas required for inclusion of a maternal treein the analysis. This resulted in a final samplesize of four maternal trees from the Blue Riverpopulation, four trees from DrummondSwamp West, and all five trees from Drum-mond Swamp East.
DNA extraction from seeds followed amodified protocol of Kang et al. (1998). Theembryo was carefully removed from the seedcoat (to ensure no maternal tissue wasincluded) and incubated at 37uC for 1 hr in400 ml of a seed DNA extraction bufferconsisting of 200 mM Tris (pH 8), 200 mMNaCl, 25 mM EDTA, and 0.5% SDS. Theembryo was then crushed in seed extractionbuffer, followed by the addition of 400 ml ofCTAB buffer. DNA extraction then followedthe protocol of Doyle and Doyle (1987),except the aqueous layer from the chloroform:isoamyl alcohol (24:1 v/v) extraction wasprecipitated in cold isopropanol rather thanethanol (Jones and Gibson 2011).
Microsatellite GenotypingNine microsatellite primer loci developedspecifically for seaside alder (Lance et al.2009) were used to determine adult andoffspring genotypes: Alma3, Alma5, Alma7,Alma11, Alma16, Alma20, Alma21, Alma25,and Alma27. Primers were fluorescently la-beled with Applied Biosystems (Foster City,California) Dye Set G5 for PCR multiplexing.The resulting microsatellite amplicons wererun on an Applied Biosystems 3130xl GeneticAnalyzer and the fragment length polymor-phisms were analyzed using Applied Biosys-tems GeneMapper 3.7 to determine genotypesfor both adults and offspring. Adult genotypesof maternal trees were determined in aprevious study examining genetic diversity
and structure within and among populations(Jones and Gibson 2011). MICRO-CHECKER2.2 (van Oosterhout et al. 2004) was used todetect microsatellite null alleles, stuttering,and large allele dropout.
Mating System AnalysisMating system analysis was conducted for theOklahoma and Georgia populations to esti-mate single- and multilocus outcrossing rates(ts and tm, respectively) at the populationslevel and for individual trees. Maximumlikelihood estimates of ts and tm were obtainedfrom maternal genotypes and offspring geno-type arrays with MLTR 3.4 using the Expec-tation-Maximization Method (Ritland 2002).The level of biparental inbreeding (sb) wasestimated by the difference between popula-tion-level estimates of tm and ts (sb 5 tm 2 ts).The inbreeding coefficient F was calculated forthe progeny arrays of maternal trees, andstatistical significance of deviations of F fromzero were tested with a x2 statistic where x2 5
F2N(a 2 1), and degrees of freedom 5 a(a 2
1)/2 where N is the sample size and a is thenumber of alleles (Li and Horvitz 1953).Multilocus correlations of paternity (rpm) werecalculated to estimate the frequency of sharedpaternity among seeds of a maternal tree.Standard error of mating system parameterswas calculated using 1,000 bootstrap repli-cates. To evaluate statistical significance ofoutcrossing and biparental inbreeding rates,95% confidence intervals were calculated todetermine whether estimates were significant-ly different from 1.00 and 0.00, respectively.Two-tailed unpaired t-tests with Welch correc-tion using GraphPad InStat v. 3.00 (Graph-Pad Software, San Diego, California) was usedto compare mating system parameters be-tween populations and subpopulations. Wecompared estimated pollen genotypes forindividual offspring to the genotypes of adulttrees in both study populations by hand toidentify potential pollen parents in the indi-vidual study populations or from surroundingpopulations in the Oklahoma subspecies.
RESULTSMating SystemsSufficient viable seed for analysis were ob-tained from 4 families and 85 offspring in theOklahoma population. Nine families and 213offspring were analyzed from Georgia. An
14 CASTANEA VOL. 77
outcrossing rate of tm 5 0.942 6 0.132(estimate 6 standard error) was estimatedfor the Oklahoma population (Table 1), anda similarly high value (tm 5 0.943 6 0.059)was estimated for the Georgia population.Mean single-locus estimates were slightlylower. Multilocus outcrossing rates for indi-vidual trees in Oklahoma ranged from 0.918to 1.076, and single-locus outcrossing ratesranged from 0.765 to 1.116 (Table 1). InGeorgia, multilocus outcrossing rates forindividual maternal trees were similarly high,ranging from 0.873 to 1.033. Single-locusoutcrossing rates were much more variableand ranged from 0.663 to 1.232. None of themultilocus estimates for populations or indi-viduals were significantly different from 1.00.
Biparental inbreeding in the Oklahomapopulation (sb 5 0.095 6 0.050) was signifi-cantly higher (p , 0.01) than in Georgia (sb 5
0.048 6 0.048), but estimates for neitherpopulation were significantly greater than0.00 (Table 1). Estimates of biparental in-breeding in Drummond Swamp West (sb 5
0.073 6 0.015) were significantly greater than0.00 and significantly higher (p , 0.01) thanDrummond Swamp East (sb 5 0.023 6 0.018).
Correlations of paternity in Oklahoma (rpm
5 0.137 6 0.094) were significantly higher (p, 0.05) than in Georgia (rpm 5 0.092 6 0.059).Within the Georgia population, correlatedpaternity in Drummond Swamp West (rpm 5
0.073 6 0.038) was significantly higher (p ,
0.01) than Drummond Swamp East (tm 2 ts 5
0.023 6 0.018, rpm 5 0.020 6 0.014). Wright’s F
was not significantly greater than zero in anypopulation, indicating that seed pools are inHardy-Weinberg equilibrium (Table 1).
Potential pollen parents for offspring siredin both Drummond Swamp East and Westcould be identified from both regions of thepopulation. In the Blue River population,most potential pollen parents could be iden-tified within the Blue River population, andoften within the immediate vicinity of thematernal tree. However, potential pollenparents for three seeds were identified fromtrees growing in the population on DelawareCreek approximately 5 km east of the BlueRiver, and potential pollen parents for eightseeds were identified from the population onPennington Creek approximately 7 km south-west of the Blue River population. T
ab
le1.
Ma
tin
gsy
stem
sp
ara
mete
rsa
nd
sta
nd
ard
err
or
(SE
)o
fO
kla
ho
ma
(OK
)a
nd
Geo
rgia
(GA
)p
op
ula
tio
nsa
Po
pu
lati
on
(in
div
idu
al)
No
.o
fFa
mil
ies
No
.o
fO
ffsp
rin
gF
(SE
)t m
(SE
)t s
(SE
)s b
(SE
)r p
m(S
E)
Okla
hom
aover
all
485
0.0
00
(0.0
21)
0.9
42
(0.1
32)
0.8
46
(0.1
03)
0.0
95
(0.0
50)
0.1
37
(0.0
94)
(OK
1)
22
1.0
47
(0.0
08)
0.9
52
(0.0
60)
(OK
2)
20
1.0
76
(0.1
01)
1.0
21
(0.0
07)
(OK
3)
23
1.0
16
(0.0
02)
1.1
16
(0.1
01)
(OK
4)
22
0.9
18
(0.0
53)
0.7
65
(0.1
06)
Geo
rgia
over
all
9213
0.0
0(0
.001)
0.9
43
(0.0
59)
0.8
95
(0.0
48)
0.0
48
(0.0
48)
0.0
92
(0.0
59)
(GA
1)
23
1.0
23
(0.0
14)
1.0
33
(0.0
66)
(GA
2)
25
0.9
60
(0.0
73)
0.8
46
(0.1
04)
(GA
3)
21
0.9
53
(0.0
40)
0.8
24
(0.0
68)
(GA
4)
25
0.9
40
(0.0
63)
0.6
63
(0.1
04)
(GA
5)
23
1.0
14
(0.0
02)
0.9
36
(0.0
85)
(GA
6)
22
0.9
31
(0.0
49)
0.8
02
(0.1
29)
(GA
7)
27
1.0
33
(0.0
04)
1.2
32
(0.0
59)
(GA
8)
24
1.0
08
(0.0
01)
0.9
19
(0.0
96)
(GA
9)
23
0.8
73
(0.1
16)
0.7
19
(0.1
58)
aF
5in
bre
edin
gco
effi
cien
t;t m
5m
ult
ilocu
sou
tcro
ssin
gra
te;
t s5
sin
gle
-locu
sou
tcro
ssin
gra
te;
s b5
rate
of
ap
pa
ren
tse
lfin
gd
ue
tob
ipa
ren
tal
inb
reed
ing;
r pm
5
mu
ltil
ocu
sco
rrel
ati
on
of
pa
tern
ity.
2012 JONES, GIBSON: MATING SYSTEM ANALYSIS OF SEASIDE ALDER 15
DISCUSSION The multilocus mating systemestimates for both study populations show thatA. maritima has a predominantly outcrossedmating system. These results are consistentwith our predictions and data from othermonoecious, wind-pollinated, woody, perenni-al species and species with highly fragmentedpopulations (Loveless and Hamrick 1984,Culley et al. 2002, Bacles et al. 2005, O’Connellet al. 2006). The estimated outcrossing rates forA. maritima are also similar to what has beenobserved in other Alnus species. Xie et al. (2002)estimated a mean outcrossing rate of tm 5 0.85for A. rubra Bong. and Bousquet et al. (1987a)estimated outcrossing at tm 5 0.95 for A. crispa
(W. Aiton) Pursh. Steiner and Gregorius (1999)identified both pre- and postzygotic incompat-ibility mechanisms, as well as inbreedingdepression acting in A. glutinosa (L.) Gaertn.,which resulted in successful seed productionfrom outcross mating events and terminatedproduction of offspring from self-fertilizationand mating among closely related individuals.These mechanisms can thereby promote thehigh outcrossing rate and population geneticdiversity that has been measured for A. mar-
itima and other Alnus species.
While both the Oklahoma and Georgiapopulations had similar outcrossing rates, aslightly higher degree of biparental inbreed-ing and correlated paternity was detected inthe Oklahoma population. These data sug-gest that maternal trees in the Blue Riverpopulation are potentially mating with re-lated individuals to a greater extent thantrees in the Drummond Swamp population.However, there was not a significant level ofbiparental inbreeding detected in either pop-ulation, and the inbreeding coefficient wasnot significantly different from 0.00 in prog-eny or the adult population (Gibson et al.2008, Jones and Gibson 2011). Althoughsome trees had single-locus outcrossing esti-mates that were considerably lower thanmultilocus estimates (Table 1), these lowersingle-locus values are likely a reflection oflower allelic variation at some loci and not ahigh degree of consanguineous matings.However, it is important to note that matingsystem estimates are from viable seeds only.Unfilled seeds cannot be used in matingsystem analysis and the high number ofunfilled seeds in some trees that resulted in
their exclusion from the study may be theresults of inbreeding that terminated seeddevelopment. Further studies are necessary todetermine the relationship among popula-tion size, inbreeding, reproductive success,and recruitment, particularly in smallerseaside alder populations.
Correlations of paternity show that pollenparents are siring multiple offspring in theprogeny of maternal trees. The differences incorrelations of paternity between the Okla-homa and Georgia populations and betweenDrummond Swamp East and West may beattributable to the presence of vegetation thatinhibits mixing and genetic homogenizationof the pollen cloud, which can have asignificant influence on the mating systemof wind-pollinated species (El-Kassaby et al.1986, Gibson and Hamrick 1991, Williams2008). Bacles et al. (2005) demonstrated thatpopulation fragments of Fraxinus excelsior L.in open landscapes tend to have greatermovement of wind-dispersed pollen thanpopulations growing in more dense vegeta-tion. This appears to be the case for A.
maritima as well. For example, in the Okla-homa population, trees grow in scatteredclumps among dense, diverse, riparian vege-tation. Comparisons of estimated pollengenotypes that sired offspring with the geno-types of individuals in the Blue River popula-tion found that potential pollen parents formany offspring were predominantly in theimmediate vicinity of a maternal tree. Theclumped distribution of seaside alder inter-spersed with other riparian vegetation likelylimits pollen movement to the immediatevicinity of the pollen parent where it wasreleased and increases their allele frequenciesin that vicinity. In contrast, potential pollenparents for seeds produced in both regions ofDrummond Swamp could be identified fromtrees growing in both Drummond SwampWest and East. The more open habitat of theDrummond Swamp population likely allowsgreater pollen movement and mixingthroughout the population. Thus, maternaltrees can receive pollen from a greaternumber of trees, and from more distant treesthroughout the population. The potentialinfluence of vegetation density on the matingsystem is also suggested by further consider-ation of mating patterns in the Georgia
16 CASTANEA VOL. 77
population. Higher correlations of paternitywere also measured in the denser stand oftrees in Drummond Swamp West as com-pared to the very open, lower density stand inDrummond Swamp East.
Although vegetation may restrict gene flow,there was indication that some seeds pro-duced by maternal trees in the Blue Riverpopulation were sired by pollen from otherpopulations. Thus, there may be sufficientpollen-mediated gene flow among popula-tions in Oklahoma such that physicallyisolated fragments are not necessarily genet-ically isolated, a key issue for conservation(Mills and Allendorf 1996). Our results areconsistent with indirect estimates by Jonesand Gibson (2011) that suggested extensivegene flow among separate Oklahoma popu-lations (Nm 5 1.93), among different regionswithin the Blue River population (Nm 5 4.45),and between Drummond Swamp East andWest (Nm 5 1.49). The potential for highlevels of gene flow in seaside alder is consis-tent with gene flow estimates of Nm 5 2.63among networks of A. hirsuta (Spach.) Rupr.populations on the Korean peninsula that areseparated from other populations by approx-imately 40 km (Huh and Huh 1999). Highergene flow of Nm 5 12.3 was also estimatedamong closely spaced populations of A. firma
Siebold & Zucc. (Huh and Huh 2001). Like-wise, extensive gene flow has also beendetected among networks of A. crispa and A.
rugosa Du Roi populations that are separatedby an average of approximately 84 km(Bousquet et al. 1987a, 1987b, 1988). Whilecomparisons of Nm across studies shouldbe interpreted with caution, our results areconsistent with results from other alder spe-cies showing the ability to experience highlevels of gene flow. Further investigations ofhow vegetation structure, landscape hetero-geneity, and isolation specifically influencethe mating system of seaside alder will requireusing a greater number of microsatellite loci,extensive mapping and genotyping of indi-vidual trees in populations, and detailedpaternity analysis to more precisely deter-mine patterns of pollen movement in thiswind-pollinated species.
Although seaside alder has a highly out-crossed mating system that produces geneti-cally diverse offspring, a critical concern for
this species is its current failure to recruitindividuals into existing populations fromseed. Seed germination is rare and successfulseedling establishment has not been ob-served in Oklahoma (Rice and Gibson 2009)or Georgia (Gibson, pers. obs.). Thus, al-though a genetically diverse seed pool isproduced, failure of seeds to germinate andestablish new individuals negates any bene-fits of sexual reproduction and renders theseed pool irrelevant. Existing populations inOklahoma and Georgia rely exclusively onasexual reproduction via rootsprouts to re-place damaged individuals, and this isprobably the case in Delmarva as well.However, when individual trees die, theyare not replaced by new seaside alders (P.Gibson, pers. obs.). This leaves seaside alderpopulations vulnerable to being replaced oroutcompeted by other species (Schrader et al.2006). Likewise, loss of individuals can resultin loss of genetic variation that may benecessary for populations to adapt andrespond to future climate change. Althougha microsatellite study of adult seaside aldertrees found no significant inbreeding in anyregional populations, there was considerablyless genetic variation in Georgia (as mea-sured by percentage polymorphic loci, aver-age number of alleles, and observed andexpected heterozygosity) than in Delmarvaor Oklahoma populations (Jones and Gibson2011). Loss of variation is expected for thislone population in Georgia that is highlyisolated from any other sources of geneticvariation. However, it may be possible toprevent such genetic decay in the remainingOklahoma and Delmarva populations.
Conservation efforts for seaside alder inparticular and other long-lived, woody peren-nial species in general should therefore focuson determining strategies to promote estab-lishment of new individuals, preserve existingpopulations, and prevent further loss ofspecies-, regional-, and population-level ge-netic variation (Gibson et al. 2008). Ourresults highlight a need for continued studiesof A. maritima utilizing a landscape geneticsapproach that integrates landscape ecology,natural disturbance regimes, reproductivebiology, and population genetic data forunderstanding the genetic dynamics of spe-cies that are the focus of conservation efforts
2012 JONES, GIBSON: MATING SYSTEM ANALYSIS OF SEASIDE ALDER 17
(Sork et al. 1999, Manel et al. 2003, Sork andSmouse 2006). For example, disturbancecaused by hurricanes, flood, fires, and otherevents should be studied to understand theirinfluence on providing safe sites for seedlinggermination and establishment. Furthermore,as climatic changes alter the distribution ofsuitable habitats for species such as seasidealder, it will be necessary to utilize ourunderstanding of their current ecology andreproductive biology that may help anticipatethe future distribution of seaside alder andenhance effective conservation planning forthis uniquely distributed species.
ACKNOWLEDGMENTS Funding for thisproject was provided from the University ofOklahoma Research Council, the Universityof Oklahoma Department of Botany-Microbi-ology, and the University of OklahomaDepartment of Zoology. Additional fundingprovided by the Garden Club of America andCenter for Plant Conservation via the Cathe-rine H. Beattie Fellowship.
LITERATURE CITEDBacles, C.F.E., J. Burczyk, A.J. Lowe, and R.A.
Ennos. 2005. Historical and contemporarymating patterns in remnant populations ofthe forest tree Fraxinus excelsior. Evolution59:979–990.
Barrett, S.C.H. and J. Kohn. 1991. The geneticand evolutionary consequences of smallpopulation size in plants: implications forconservation. p. 3–30. In: D. Falk and K.E.Holsinger (eds.). Genetics and conservationof rare plants. Oxford University Press, NewYork, New York.
Bittencourt, J.V.M. and A.M. Sebbenn. 2007.Patterns of pollen and seed dispersal in asmall, fragmented population of the wind-pollinated tree Araucaria angustifolia insouthern Brazil. Heredity 99:580–591.
Bousquet, J., W.M. Cheliak, and M. Lalonde.1987a. Allozyme variability in naturalpopulations of green alder (Alnus crispa) inQuebec. Genome 29:345–352.
Bousquet, J., W.M. Cheliak, and M. Lalonde.1987b. Genetic differentiation among 22populations of green alder (Alnus crispa) incentral Quebec. Canad. J. Forest Res.17:219–227.
Bousquet, J., W.M. Cheliak, and M. Lalonde.1988. Allozyme variation within andamong mature populations of speckledalder (Alnus rugosa) and relationships withgreen alder (A. crispa). Amer. J. Bot. 75:1678–1686.
Cardoso, M.A., J. Provan, W. Powell, P.C.G.Ferreira, and D.E. De Oliveira. 1998. Highgenetic differentiation among remnantpopulations of the endangered Caesalpiniaechinata Lam. (Leguminosae-Caesalpinioi-deae). Molec. Ecol. 7:601–608.
Chen, Z. and J. Li. 2004. Phylogenetics andbiogeography of Alnus (Betulaceae) inferredfrom sequences of nuclear ribosomal DNAITS region. Int. J. Pl. Sci. 165:325–335.
Coates, D.J., J.F. Sampson, and C.J. Yates.2007. Plant mating systems and assessingpopulation persistence in fragmented land-scapes. Austral. J. Bot. 55:239–249.
Culley, T.M., S.G. Weller, and A.K. Sakai. 2002.The evolution of wind pollination in angio-sperms. Trends Ecol. Evol. 17:361–369.
Davis, M.B. and R.G. Shaw. 2001. Range shiftsand adaptive responses to quaternary cli-mate change. Science 292:673–679.
Delaware Natural Heritage Program. 2003.Rare vascular plants of Delaware (http://www.dnrec.state.de.us/fw/rareplant03.pdf)(downloaded 25 August 2006). Dela-ware Natural Heritage Program, Smyrna,Delaware.
Doyle, J.J. and J.L. Doyle. 1987. A rapid DNAisolation procedure for small quantities offresh leaf tissue. Phytochem. Bull. 19:1–15.
El-Kassaby, Y.A., J. Parkinson, and W.J.B.Devitt. 1986. The effect of crown segmenton the mating system in Douglas-fir (Pseu-dotsuga menziesii (Mirb.) Franco) seed or-chard. Silvae Genet. 35:149–155.
Ellstrand, N.C. and D.R. Elam. 1993. Popula-tion genetic consequences of small popula-tion size: implications for plant conserva-tion. Annual Rev. Ecol. Syst. 24:217–242.
Georgia Division of Natural Resources. 2006.Special concern plant species in Georgia(http://georgiawildlife.dnr.state.ga.us/content/specialconcernplants.asp) (accessed 26 Octo-ber 2006). Georgia Division of Natural Re-sources, Social Circle, Georgia.
18 CASTANEA VOL. 77
Gibson, J.P. and J.L. Hamrick. 1991. Geneticdiversity and structure in Pinus pungens(Table Mountain Pine) populations. Canad.J. Forest Res. 21:635–642.
Gibson, J.P., S.A. Rice, and C.M. Stucke. 2008.Comparison of population genetic diversitybetween a rare, narrowly distributed speciesand a common, widespread species of Alnus(Betulaceae). Amer. J. Bot. 95:588–596.
Gibson, J.P. and N.T. Wheelwright. 1995.Genetic structure in a population of a tropicaltree, Ocotea tenera (Lauraceae): influence ofavian seed dispersal. Oecologia 103:49–54.
Hamrick, J.L. and M.J.W. Godt. 1990. Allo-zyme diversity in plant species. p. 43–61. In:A.H.D. Brown, M.T. Clegg, A.L. Kahler, andB.S. Weir (eds.). Plant population genetics,breeding, and genetic resources. Sinauer,Sunderland, Massachusetts.
Huh, M.K. and H.W. Huh. 1999. Geneticdiversity and population structure of Alnushirsuta (Betulaceae) in Korea. J. Pl. Res.112:437–442.
Huh, M.K. and H.W. Huh. 2001. Geneticdiversity and phylogenetic relationships inalder, Alnus firma, revealed by AFLP. J. Pl.Biol. 44:33–40.
IUCN. 2010. IUCN red list of threatened species.Version 2010.1 (www.iucnredlist.org) (down-loaded 9 April 2010). International Union forConservation of Nature, Gland, Switzerland.
Jones, J.M. and J.P. Gibson. 2011. Populationgenetic diversity and structure within andamong disjunct populations of Alnus mar-itima (seaside alder) using microsatellites.Conservation Genet. 12:1003–1013.
Jump, A.S. and J. Penuelas. 2006. Geneticeffects of chronic habitat fragmentation ina wind-pollinated tree. Proc. Natl. Acad. Sci.U.S.A. 103:8096–8100.
Kang, H.W., Y.G. Cho, U.H. Yoon, and M.Y.Eun. 1998. A rapid DNA extraction methodfor RFLP and PCR analysis from a single dryseed. Pl. Molec. Biol. Reporter 16:1–9.
Karron, J.D. 1991. Patterns of genetic varia-tion and breeding systems in rare plantspecies. p. 87–98. In: D. Falk and K.E.Holsinger (eds.). Genetics and conservationof rare plants. Oxford University Press, NewYork, New York.
Kelly, A.E. and M.L. Goulden. 2008. Rapidshifts in plant distribution with recentclimate change. Proc. Natl. Acad. Sci.U.S.A. 105:11823–11826.
Lance, S.L., K.L. Jones, C. Hagen, T.C. Glenn,J.M. Jones, and J.P. Gibson. 2009. Develop-ment and characterization of nineteenpolymorphic microsatellite loci from sea-side alder, Alnus maritima. ConservationGenet. 10:1907–1910.
Li, C.C. and D.G. Horvitz. 1953. Some meth-ods of estimating the inbreeding coefficient.Amer. J. Hum. Genet. 5:107–117.
Little, L.E. 1975. Our rare and endangeredtrees. Amer. Forests 81:16.
Loveless, M.D. and J.L. Hamrick. 1984. Eco-logical determinants of genetic structure ofplant populations. Annual Rev. Ecol. Syst.15:65–95.
Manel, S., M.K. Schwartz, G. Luikart, and P.Taberlet. 2003. Landscape genetics: com-bining landscape ecology and populationgenetics. Trends Ecol. Evol. 18:4, 189–197.
Maryland Department of Natural Resources.2004. Current and historical rare, threatenedand endangered species (http://www.dnr.state.md.us/wildlife/rte/rte04worc.pdf) (down-loaded 25 August 2006). Maryland Depart-ment of Natural Resources, Annapolis, Mary-land.
Mills, L.S. and F.W. Allendorf. 1996. The one-migrant-per-generation rule in conserva-tion and management. Conservation Biol.10:1509–1518.
NatureServe. 2010. NatureServe explorer: anonline encyclopedia of life. Version 7.1(http://www.NatureServe.org/explorer) (ac-cessed 18 August 2010). NatureServe, Ar-lington, Virginia.
O’Connell, L.M., A. Mosseler, and O.P. Ra-jora. 2006. Impacts of forest fragmentationon the mating system and genetic diversityof white spruce (Picea glauca) at the land-scape level. Heredity 97:418–426.
Oklahoma Biological Survey. 2010. Workinglist of rare Oklahoma plants (http://www.biosurvey.ou.edu/download/heritage/plants0503.pdf) (downloaded 10 April 2010). Okla-homa Biological Survey, Norman, Oklahoma.
2012 JONES, GIBSON: MATING SYSTEM ANALYSIS OF SEASIDE ALDER 19
Rice, S.A. and J.P. Gibson. 2009. Is seedlingestablishment very rare in the Oklahomaseaside alder, Alnus maritima ssp. oklahomen-sis? Oklahoma Native Plant Record 9:59–63.
Ritland, K. 2002. Extensions of models forthe estimation of mating systems using nindependent loci. Heredity 88:221–228.
Schrader, J.A. and W.R. Graves. 2000. Timingof seed dispersal may limit the reproductivesuccess of Alnus maritima. Castanea 65:69–77.
Schrader, J.A. and W.R. Graves. 2002. Infra-specific systematics of Alnus maritima (Betu-laceae) from three widely disjunct prove-nances. Castanea 67:380–401.
Schrader, J.A., W.R. Graves, S.A. Rice, and J.P.Gibson. 2006. Differences in shade toler-ance help explain varying success of twosympatric Alnus species Int. J. Pl. Sci.167:979–989.
Sork, V.L., J. Nason, D.R. Campbell, and J.F.Fernandez. 1999. Landscape approaches tohistorical and contemporary gene flow inplants. Trends Ecol. Evol. 14:219–224.
Sork, V.L. and P.E. Smouse. 2006. Geneticanalysis of landscape connectivity in treepopulations. Landscape Ecol. 21:821–836.
Steiner, W. and H.-R. Gregorius. 1999. Incom-patibility and pollen competition in Alnus
glutinosa: evidence from pollination exper-iments. Genetica 105:259–271.
Stibolt, V.M. 1981. The distribution of Alnus
maritima Muhl. ex Nutt. (Betulaceae). Cas-tanea 46:195–200.
Thuiller, W., C. Albert, M.B. Araujo, P.M.Berry, M. Cabeza, A. Guisan, T. Hickler,G.F. Midgley, J. Paterson, F.M. Schurr, M.T.Sykes, and N.E. Zimmermann. 2008. Pre-dicting global change impacts on plantspecies distributions: future challenges. Per-spect. Pl. Ecol. 9:137–152.
van Oosterhout, C., W.F. Hutchinson, D.P.M.Wills, and P. Shipley. 2004. MICRO-CHECKER:software for identifying and correcting geno-typing errors in microsatellite data. Molec.Ecol. Notes 4:535–538.
Williams, C. 2008. Aerobiology of Pinus taeda
pollen clouds. Canad. J. Forest Res. 38:2177–2188.
Xie, C.-Y., Y.A. El-Kassaby, and C.C. Ying.2002. Genetics of red alder (Alnus rubra
Bong.) populations in British Columbia andits implications for gene resources manage-ment. New Forest 24:97–112.
20 CASTANEA VOL. 77
