Pentanthera Webs: Interspeci!c and Interploid Hybridization among Sympatric Azaleas in the Southern Appalachian Mountains
Kimberly Shearer, !omas G. this bald suggests that the present azalea Ranney, Ron Miller, and Clarence species in the Southeast are likely to be re- Towe cent products of cycles of migration and Mountain Crop Improvement Lab, interaction attending the as many as 20 Department of Horticultural glaciations and warm periods since the be- Science, Mountain Horticultural ginning of the Pleistocene. Our contem- Crops Research and Extension porary list of recognized azalea species that Center, North Carolina State Kimberly Shearer Thomas Ranney seem so distinct and certain may simply be University, Mills River, North a momentary snapshot of a complex and Carolina rapidly changing evolutionary web that is the Pentanthera. ummary SModern evolutionary research sug- that famous swarm. !ere were tetraploids Introduction gests that new species often arise rapidly keying out to R. calendulaceum, some with Beginning with Charles Darwin (Darwin from hybridization and chromosome unusual variations (e.g., fragrance), within 1859), the evolutionary process typically doubling, augmenting the slow, divergent nearby woodlands. !e traits of several has been illustrated as a tree of life with processes originally detailed by Darwin. of the samples from Wayah/Wine Spring radiating limbs that branch and diverge Relationships between kindred species gave evidence for genetic exchanges be- but never reconnect. !is concept is are thus best represented not by a simple tween the diploid and tetraploid species; rooted in the notion of reproductively branching candelabra or tree, as pictured one triploid, such as might serve as a two- isolated populations accumulating genetic in our old biology texts, but by a complex way bridge between the azalea species, variation over time. Subject to natural web of exchanges and ploidy variations. was in fact found. In addition, triploids selection, these populations become Such complexities seem especially evi- had previously been discovered from R. increasingly distinctive as they gradually dent in the highly compatible and multi- calendulaceum-R. periclymenoides hybrids evolve along a one-way course without ploidal Southeastern deciduous azaleas, in northwestern South Carolina. !us intersections. Current research provides especially on those spectacular and much- mechanisms for gene exchange between ample evidence that this tree-of-life visited mountaintops called “balds.” We diploid and tetraploid species do indeed illustration is an oversimpli#cation at have conducted a survey of 92 samples exist. Examination of the complex dip- best (Arnold 2006; Arnold and Larson, taken from the azalea swarm on the bare loids of Gregory was still more suggestive. 2004). !e process of evolution is more top and adjacent woodlands of Gregory Flower colors, odors, and forms evince aptly depicted as a reticulate web of Bald within Great Smokies National Park. origins in R. cumberlandense and R. arbo- genetic exchange integrating periodic Flow cytometry was used to determine rescens and possibly R. viscosum. Diploid hybridization among nascent “species.” ploidy and published component traits of plants consistent with R. cumberlandense Interspeci#c hybridization and intro- recognized species were used to elucidate at this site could not, based on "ower sizes gression are common among plants and mating interactions at that highly diverse and forms, be distinguished from tetra- are a signi#cant evolutionary mechanism site. A similar study was undertaken with ploid R. calendulaceum. !e lack of dif- that can give rise to new evolutionary 12 samples from along the interfaces be- ferentiating morphological traits between lineages and species (Arnold 1997, 2006; tween the Rhododendron calendulaceum R. cum-berlandense and R. calendulaceum Barrier et al. 1999; Grant et al. 2005; (tetraploid) and R. arborescens (diploid) and similarity in average chromosome size Rieseberg and Carney 1998; Soltis et colonies on Wayah/Wine Spring Bald, to suggests that tetraploid R. calendulaceum al. 2009). By some estimations, at least the south of the Park. Cytometry revealed at this site is primarily derived from the 25% of plant species hybridize naturally no tetraploids within the open bald area of diploid R. cumberlandense with limited ge- (Mallet 2005) and as many as 50-70% Gregory, belying previous suggestions that netic infusions from other species. More- of all "owering plants are of hybrid origin R. calendulaceum is directly involved in over, the indeterminacy of species upon (Ellstrand et al. 1996; Rieseberg 1997).
JOURNAL AMERICAN RHODODENDRON SOCIETY 187 !is process of introgressive hybridization and Vorsa 1993; Ramsey and Schemske Although these grassy balds have certainly can a$ord evolutionary bene#ts by 1998). !is reproductive barrier between been disturbed and impacted by human enhancing genomic diversity. !rough the cytotypes can foster sympatric divergence. beings (Lindsay 1976), they contain development and combination of novel Yet little is known about the potential hybrid zones of deciduous azaleas and adaptations, introgressive hybridization for in situ interploid gene "ow among provide model sites to study ongoing can enhance #tness, colonization success, coexisting taxa of di$erent cytotypes. genetic exchange. and adaptive radiation in new and Hybrid zones can provide an oppor- !e species found on these balds are changing environments (Arnold 2006; tunity to study the degree of reproductive characterized and may be di$erentiated Grant et al. 2005; Lewontin and Birch isolation and the presence of interspe- in the following ways (Galle 1987; Kron 1966; Lindqvist et al. 2003; Rieseberg and ci#c and interploid crossing within mixed 1993; Luteyn et al. 1996; Willingham, Jr. Carney 1998; Soltis et al. 2009). populations. Pentanthera azaleas native to 1976) (also see Table 1): Rhododendron species of the sub-sec- the Southern Appalachian Mountains of- tRhododondron arborescens is known tion Pentanthera are known for their wide- ten form hybrid zones or “swarms” with as the sweet azalea due to the sweet clove spread natural hybridization, including unusual phenotypic diversity indicative of or heliotrope-like fragrance. Flowers are over 18 di$erent combinations that some- natural hybridization (Galle 1987; Kron white, sometimes with a pink or reddish times involve multiple species (Millais 1993; Rehder 1921; Skinner 1955, 1961; blush, 4.0-5.2 cm across, with or without 1924; Skinner 1955, 1961; Galle 1968; Towe 2004). A number of studies have a yellowish blotch. !e transition from Leach 1958). Leach (1958) remarked that con#rmed interspeci#c hybridization and tube to corolla is gradual; sepals typically “hybridity is the most conspicuous single introgression among these azaleas, but have a fringed margin. Pistil and stamens feature of the entire Azalea population of these have been limited to diploid cyto- are characteristically reddish. our Southern mountains.” Skinner (1961) types (King 2000; Kron et al. 1993). tRhododendron calendulaceum, the also emphasized what he considered to be A well-known hybrid zone of "ame azalea, is named for its brilliant the reticulate nature of the evolutionary azaleas is located on Gregory Bald in "ower color which can range from yellow lineages within this section. the Great Smoky Mountains National to orange to red. Flowers are often larger !e development of polyploids is Park on the Tennessee-North Carolina than the "owers of other species, 4.0- also a fundamental evolutionary process border near Cades Cove. Based on 6.5 cm across, and have a blotch on the in "owering plants. Most angiosperms morphology and the overlapping ranges upper lobe. !ere is an abrupt transition are believed to have undergone whole of numerous species, this swarm has been from tube to corolla, no substantial genome duplication events, often followed thought to represent a complex hybrid fragrance, and typically a fringed sepal by rediploidization, throughout their zone among Rhododendron arborescens margin. Flowers (2n=4x=52) often bloom evolutionary history (Soltis et al. 2009). (Pursh), Rhododendron cumberlandense before leaf emergence (though highly Polyploids have numerous traits that can (E.L. Braun) Copeland, Rhododendron variable). Rhododendron calendulaceum contribute to successful speciation and calendulaceum (Michx.) Torr., and is the only tetraploid species of the four radiation, including a greater number Rhododendron viscosum (L.) Torrey (Hyatt taxa considered here (Jones et al. 2007; Li of potential alleles, greater potential 2001). Potential hybridization events in 1957; Sax 1930). heterozygosity, tolerance of deleterious this population are of particular interest tRhododendron cumberlandense (R. mutations, novel gene expression(s), on account of the complex genetic bakeri), the Cumberland azalea, closely and enzymatic multiplicity (Adams and interactions that may be occurring resembles R. calendulaceum but is diploid. Wendel 2005; Comai 2005; Hegarty between species that vary in ploidy level. !is species has been described as having and Hiscock 2008; Osborn et al. 2003; Of the four species found in that region, smaller "owers, 3.8-4.5 cm across, a bristly Soltis et al. 2003; Soltis and Soltis 1993). R. arborescens, R. cumberlandense, and sepal margin, and it typically blooms after !e formation of polyploids can also R. viscosum are diploid (2n = 2x = 26), leaves expand. !e pedicel is typically contribute to the development of more while R. calendulaceum is tetraploid (2n eglandular. abrupt speciation as a result of reproductive = 4x = 52) (Jones et al. 2007; Sax 1930). tRhododendron viscosum (var. monta- isolation. Once tetraploids arise in a Approximately 47 kilometers to the SE num), the sticky azalea, has fragrant white population, they generally hybridize of Gregory Bald, Wayah Bald and Wine "owers, 2.5-4.0 cm across, a glandular/vis- readily with other tetraploids. Crosses Spring Bald are located in the Nantahala cid tube, a gradual transition from tube to with diploids are typically less successful. National Forest in North Carolina. Both corolla, and a strong spicy fragrance. Rho- Successful crosses between tetraploids Wayah Bald and Wine Spring Bald host dodendron viscosum can be confused with and diploids typically yield triploids with sympatric populations of R. calendulaceum, R. arborescens as they both have fragrant low or nonexistent fertility (Ehlenfeldt R. arborescens, and their potential hybrids. white "owers; however, unlike the R. arbo-
188 FALL 2012 Table" > %'& %"% 1. Morphological " ') '(% %*' characters of four *%*(/" (6deciduous azaleas (Rhododendron 27$&*)) sp.) and + .' putative ((#&" )' %'.0. $ hybrids sampled at Gregory, Wayah $ and&' $ "2Winespring Bald.
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
?> (Table contined on next page.) rescens, R. viscosum does not typically have distribution (Kron et al. 2007). bags with moist paper towels and were the tell-tale, red pistil and stamens and is a !e objective of this study was to transported in coolers. Photographs and low, stoloniferous shrub. utilize morphological and genome size data GPS coordinates were taken for each In order to better di$erentiate to determine the presence of interspeci#c plant to allow for follow-up research. these species (e.g., R. calendulaceum and and interploid hybridization in selected Plant locations and cytotype distributions R. cumberlandense) and to understand hybrid zones of Pentanthera azaleas in the were mapped (Google Earth, Google interploid hybridization, it is essential to Southern Appalachian Mountains. Inc., Mountain View, Calif.). Additional know the ploidy levels of the plants in samples were collected from selected question. Cytological determination of Materials and Methods plants and pressed for herbarium vouchers chromosome numbers and ploidy levels Flower and foliage samples were collected that are now maintained at the NC State for Rhododendron is notoriously di%cult from 92 plants growing on and around Herbarium, Raleigh, N.C. (K. Shearer, because of the small chromosome size. Gregory Bald and 12 samples from Wayah 1-38). Morphological characters, listed However, the recent development of "ow Bald and Wine Spring Bald in mid-June in Table 1, were determined. Presence cytometry provides for rapid measurement 2011. Samples were collected throughout of fragrance was con#rmed by three of genome size (DNA content) and the sites in an attempt to represent a individuals. associated ploidy levels and has been used broad range of phenotypes that were Genome sizes of all collected samples extensively for Rhododendron (Jones et al. present including plants that appeared were determined using "ow cytometry. 2007). !is method allows for rapid and to have hybrid phenotypes. Samples Approximately 1 cm² of tissue was #nely accurate screening studies of cytotype for cytometric analyses were placed in chopped in a petri dish with 400 µL of
JOURNAL AMERICAN RHODODENDRON SOCIETY 189 Table 1 continued.
=>A8 >2C=G=2=? ?- '$ @?2>A ?E2@D =>B8 @B9@>2?>F30 >2C?G=2=? ?- ) ""%, AE2=C B>2ED E@9B>2EED3 =>C8 @B9@>2?=A30 >2C>G=2=? ?- '$ @@2== @D2A? E@9B>2FBE3 =>D8 @B9@>2>FD30 >2CBG=2=> ?- ) $! ""%, @D2FA @C2>@ $!5' 4 $!5 E@9B>2FBF3 , ) =>E @B9@>2>F>30 >2CAG=2=@ ?- '$ $! A>2AC AD2=@ '$ 5& $! E@9B>2FC=3 =>F8 @B9@>2>DF30 >2C?G=2=@ ?- ""%, @A2?? A>2B@ ""%, E@9B>2FBA3 =?=8 @B9@>2>DF30 >2C?G=2==> ?- ) ""%, @E2AE @E2A> ) E@9B>2FBA3 =?>8 @B9@>2>CF30 >2C?G=2=? ?- '$ ""%, ?E2=B ?E2DA E@9B>2FC@3 =?? @B9@>2>DE30 >2C@G=2=? ?- $! '$ A=2=A B=2E= E@9B>2FC>3 =?@ @B9@>2>DE30 >2C@G=2=A ?- '$ @F2BA A?2?F E@9B>2FCD3 ""%, =?A8 @B9@>2>E>30 >2C?G=2=@ ?- $! ""%, A=2DA AC2DC $! E@9B>2FDA3 =?B8 @B9@>2>EA30 >2C?G=2=? ?- ) ""%, A@2BA @D2?F E@9B>2FD>3 =?C8 @B9@>2>FA30 >2C?G=2=@ ?- '$ @C2DA A@2>A E@9B?2=@E3 =?D8 @B9@>2>FC30 >2C@G=2=A ?- $! @F2>A AD2>E E@9B?2=@E3 =?E >2C?G=2=? ?- '$ ""%,5 @?2=D @B2>D %) $!5' %'$ =?F @B9@>2?A=30 >2C>G=2=A ?- ) ""%, @E2B@ @B2=E %) 5 E@9B>2FAE3 & $! =@= @B9@>2?A=30 >2C=G=2=A ?- '$ @D2?> A?2EA E@9B>2FA=3 =@> @B9@>2?A=30 >2C>G=2=? ?- '$ @B2B@ @E2E> $!4 E@9B>2FAE3 =@? @B9@>2??E30 @2?>G=2=@ A- '$ A?2?> AF2DD '$ E@9B?2@FB3 =@@ @B9@>2??F30 @2?=G=2=? A- '$ A?2F@ AF2AD ""%, E@9B?2@FA3 ==>8 @B9@>2?BA30 >2C=G=2=? ?- $! ""%,4 A=2AA B?2>@ E@9B>2ECA3 '$ ==?8 @B9@>2?C@30 >2C>G=2=B ?- ) ""%, '! @D2>A A=2DD ) 5 E@9B>2EDD3 . ""%, ' 4 ) ==@8 @B9@>2?CC30 >2CCG=2=@ ?- $! '$ 4 @D2== AA2== $! E@9B>2ECE3 . ""%, ==A8 @B9@>2?DC30 >2C=G=2=A ?- $! '$ @@2E@ @?2@E
?? (Table 1 continued on next page.) nuclei extraction bu$er (CyStain UV while the remaining nine were tetraploid but lacked a well-de#ned blotch (e.g., Precise P Nuclei Extraction Bu$er, Partec, (Table 1). No triploids were identi#ed A001, A003, A009, A012, A013, A016, Münster, Germany) using a sharp razor at this site. Plants that were sampled in A018, A019, A023, A026, B016, B018, blade. !e suspensions were then #ltered open areas at the top of the bald were B020, B021, B024, B032, B033, B036, through 50 µm #lters and nuclei were exclusively diploids (Fig. 1; Google Earth B037, B038, B042, B044, B046, B048, stained with 1600 µL 4’, 6-diamidino- Map, http://www.ces.ncsu.edu/"etcher/ B052, B054), a trait some authors 2-phenylindole (DAPI) staining bu$er mcilab/publications/pentantheramap. consider to be a primary characteristic of (CyStain UV Precise P Staining Bu$er, kmx. You must #rst download Google R. cumberlandense (Kron 1993). Other Partec). Relative genome sizes were Earth software to view this map: http:// plants were largely consistent with R. determined using a "ow cytometer www.google.com/earth/index.html. Click arborescens (e.g., A015, A025, B031) and (Partec PA-II, Partec) using Pisum sativum on pins to show images of each plant.) Li included a yellow blotch. We did not ‘Ctirad’ (8.75 pg) (Doležel et al. 1998) as (1957) completed chromosome counts on observe plants consistent with R. viscosum, an internal standard. Ploidy levels were eight plants from the same location and but they have been reported as present in determined based on genome size as also found them to all be diploid. the past (Kron 1993). determined by Jones et al. (2007). A number of diploid species were Many of the diploids displayed mor- clearly represented on the bald. Many phological characters that were consistent Results and Discussion plants had phenotypes consistent with R. with hybrids. For example, diploid plants Gregory Bald (A & B Group) cumberlandense (e.g., A006, A021, B006, with pink/peach "owers, with or without Of the samples collected on and B009). !ere were also many plants that a blotch, sometimes with fragrance, and around Gregory Bald, 83 were diploid, were consistent with R. cumberlandense sometimes with a red pistil and stamens
190 FALL 2012 Table 1 continued.
E@9B>2EA=3 ==B8 @B9@>2?DA30 >2C=G=2=A ?- '$ , ) @@2C= A=2== E@9B>2E@>3 )' ) %$( ==C8 @B9@>2?EF30 >2C?G=2==> ?- '$ '$ @B2CE @F2>E E@9B>2E@=3 ==D8 @B9@>2?BE30 >2C>G=2===> ?- ) ""%, A>2?D @F2?E 4 ) E@9B>2DF@3 ==E8 @B9@>2?@F30 @2>@G=2=F A- ) ""%, A>2BB A?2?? ""%, E@9B>2D>?3 $!"*( ==F8 @B9@>2?@A30 >2BCG=2=@ ?- '$ @@2F? AA2BD E@9B>2DB=3 =>=8 @B9@>2?BA30 >2C@G=2=? ?- $! '$ @D2FB A?2F= $!4 E@9B>2E=C3 =>>8 @B9@>2?BA30 >2C@G=2=> ?- ) ""%, A=2CD @?2@= 4 ) E@9B>2E=A3 =>?8 @B9@>2?AC30 >2C@G=2=> ?- $! ""%, '$ @E2>F AF2F@ ""%,4 ) E@9B>2E=C3 =>@8 @B9@>2??D30 >2C@G=2=> ?- $!, ) . ""%, A=2@F A@2D> ) $! E@9B>2E?>3 ) )' & ( =>A8 @B9@>2?>C30 >2BFG=2=A ?- '$ @E2=E B@2>B '$ E@9B>2EB>3 ""%, =>B @B9@>2?=D30 >2BEG=2=A ?- @D2@E @?2@D E@9B>2EBA3 =>C @B9@>2?BD30 >2BEG=2=? ?- '$ @E2DB @@2=@ E@9B>2FA@3 =>D @B9@>2?C@30 >2CEG=2=? ?- $! ""%, @@2AB @B2>F $! E@9B>2FBC3 =>E @B9@>2??D30 >2BBG=2==@ ?- @>2A> @@2CD E@9B>2FF=3 =>F @B9@>2?>C30 @2??G=2=@ A- '$ @F2?F A>2A? '$ 4 E@9B?2=@=3 ""%, =?= @B9@>2?=F30 >2CAG=2=A ?- '$ @A2B? A?2EA E@9B?2=@B3 =?> @B9@>2>F?30 >2C@G=2=@ ?- '$ $!5' E@9B?2=A@3 =?? @B9@>2>F?30 >2C@G=2=? ?- '$ AA2F= B=2=A E@9B?2=AB3 =?@ @B9@>2>D@30 >2CBG=2=? ?- '$ @C2CD A@2?D E@9B?2=A@3 =?A @B9@>2>EC30 >2CAG=2=C ?- A@2?C A?2>F E@9B?2=@>3 =?B @B9@>2>EA30 >2C@G=2=A ?- '$ @C2DB AA2@> E@9B?2=?F3 =?C @B9@>2>DC30 >2C=G=2=A ?- '$ @C2DF ?D2?> 4'$
?@ (Table 1 continued on next page.) were common (e.g., A002, A004, A005, pistil) and B011 (white "owers and no common trait in R. calendulaceum (Kron A007, A011, A022, A024, A029, A030, fragrance), are also most likely advanced 1993). Two tetraploids (B019 and B056) B004, B010, B013, B015, B017, B022, hybrids between R. arborescens and R. were noteworthy in that they had pheno- B027, B029, B030, B040, B041, B044, cumberlandense. Evidence for hybrids types consistent with R. calendulaceum but B045, B049, B051). Based on morphol- involving other species was di%cult to were also fragrant. !ese plants appear to ogy, these are most likely hybrids between discern based on morphology. !ese ob- represent introgression of fragrance from a R. arborescens and R. cumberlandense. servations support the premise that there diploid species (e.g., R. arborescens) into R. Plants with very similar appearance were is considerable interspeci#c hybridization calendulaceum. discovered in Vogel State Park at Neel among diploid species on Gregory Bald, Without considering genome size and Gap, Georgia in 1934 and were named R. most commonly between R. arborescens ploidy level, R. cumberlandense and R. ca- furbishii (Lemmon 1941). Leach (1958) and R. cumberlandense (Galle 1987; Kron lendulaceum were virtually indistinguish- later recreated this general phenotype by 1993). able at this site. Various traits that have hybridizing R. arborescens and R. cumber- Tetraploids were only found in un- been presented as key distinguishing char- landense and concluded that R. furbishii derstory conditions in areas surround- acteristics, including "ower size and sepal was really an interspeci#c hybrid. Leach ing and below the open bald, often in margins (Table 1), were variable within (1958) further speculated that blotched association with diploids. Many of these both cytotypes. Both taxa were bloom- R. arborescens may be the result of intro- tetraploids were consistent with R. ca- ing well after leaves had emerged at this gression from R. cumberlandense. Other lendulaceum (e.g., A032, A033, B008, site and displayed a range of "ower colors plants, including B002 (yellow "owers, B053, B055, B058, B059), though none from yellow to dark red. In a more exten- no fragrance, but with red stamens and of these plants had distinctive blotches, a sive survey of R. calendulaceum and R.
JOURNAL AMERICAN RHODODENDRON SOCIETY 191 Table 1 continued.
E@9B?2=>A3 ""%, =?D @B9@>2>F>30 >2C>G=2=@ ?- $! ""%, @@2B@ @@2AE $! E@9B?2==C3 =?E @B9@>2>F>30 >2BEG=2=C ?- @>2>F @D2CB E@9B?2==>3 =?F @B9@>2>F=30 >2C@G=2=? ?- $! ""%, A?2=@ AA2>C $!4 E@9B>2FE?3 =@= @B9@>2>ED30 >2CAG=2=@ ?- $! '$ @E2FD A=2D@ $!5' E@9B>2FDA3 =@> @B9@>2>EA30 >2CAG=2==? ?- ) ""%, A?2B@ AA2FD ) 5' E@9B>2FDA3 =@? @B9@>2?>>30 >2CAG=2=> ?- @>2@= @C2BB E@9B>2FB?3 =@@ @B9@>2?>>30 >2CAG=2=> ?- $!'$ @A2== @@2>= E@9B>2F?C3 =@A @B9@>2?=B30 >2C?G=2=C ?- $! '$ @@2A= @B2?= ) 5& $! E@9B>2F@=3 =@B @B9@>2?=C30 >2C>G=2=? ?- ) ""%, @?2== ?E2C= ) 5 E@9B>2F?A3 & $!4 =@C @B9@>2??F30 >2CDG=2=E ?- '$ @B2BC A@2?C $!5' E@9B>2EFE3 =@D @B9@>2?AE30 >2CAG=2=? ?- '$ ?F2AD @>2E= $!5 E@9B>2F?A3 =@E @B9@>2?@>30 >2CBG=2=A ?- @E2A> A?2@? E@9B>2F=>3 =@F @B9@>2??D30 >2C=G=2=> ?- $!'$ @A2=C B=2?C E@9B>2EED3 =A= @B9@>2??D30 >2C@G=2=> ?- $! A?2DC AD2F> E@9B>2EE=3 =A> @B9@>2??C30 >2C?G=2=> ?- '$ @@2=F @D2@C '$ E@9B>2EE@3 ""%, =A? @B9@>2?@=30 >2C>G=2=@ ?- '$ @B2FB B>2DD E@9B>2EDD3 =A@ @B9@>2??E30 >2C>G=2==@ ?- '$ @?2AD @@2@E E@9B>2EDF3 =AA @B9@>2?@D30 >2C>G=2=A ?- $! '$ @>2AE @F2BA $!)% E@9B>2ECE3 =AB @B9@>2?A>30 >2C=G=2=> ?- ""%, @@2D@ @A2?C E@9B>2EC>3 =AC @B9@>2?AE30 >2BFG=2==A ?- $! @@2E= A>2?C $! E@9B>2ECC3 =AD @B9@>2?AD30 >2BFG=2=? ?- $! ""%, @E2D= @C2=@ ) )% E@9B>2EBB3 $!4 =AE @B9@>2?B>30 >2CAG=2=A ?- '$ @=2?? @>2@@ E@9B>2EBE3 =AF @B9@>2?BD30 >2C?G=2=? ?- $! @E2FA @E2@A ""%,)% (Table 1 continued on next page.) ?A cumberlandense in the Nantahala Moun- Kron 1993). However, more recent work we suggest that R. calendulaceum there are tains of N.C., Willingham, Jr. (1976) by Jones et al. (2007) using "ow cytometry primarily derived from R. cumberlandense, found it was di%cult to separate the two showed that the base genome size (1Cx) with limited introgression from other species in the #eld—their "owering pe- for Pentanthera azaleas is highly conserved species, i.e., the two species represent a riod overlapped—and the cited distin- and that there is no discernible di$erence primarily autopolyploid series. However, it guishing characters (Bowers 1968; Galle in average chromosome size between R. is not unlikely that elsewhere tetraploid R. 1968; Lemmon 1937) were of little value cumberlandense, R. prinophyllum, and R. calendulaceum-like plants have arisen, and to the point of making #eld identi#cation calendulaceum. King (1977) found, in a continue to arise, through independent di%cult or impossible. Perhaps introgres- limited sampling, that R. cumberlandense genome duplication events involving sion between these two species in Western and R. calendulaceum varied slightly R. cumberlandense by itself or in various North Carolina has made it particularly in "avonoid constituents; yet his combinations with other diploid species. di%cult to clearly separate them there. phylogenetic analyses, based on "avonoid !is scenario would certainly contribute to !ere has been considerable debate constituents, showed them to be closely the great range of morphological variation and discussion on the origin of the allied. Phylogenetic analysis based on noted in R. calendulaceum (Willingham tetraploid, R. calendulaceum. Li (1957) morphological, phenological, and chemical 1973; Kron 1993). indicated that R. calendulaceum had two characters also show R. cumberlandense distinct sizes of chromosomes, suggesting and R. calendulaceum to be very similar Wayah Bald (W1-W4) it was an allotetraploid hybrid, possibly (Kron 1993). Considering these two Of the samples collected at Wayah between R. cumberlandense and R. species are virtually indistinguishable Bald, one plant was a tetraploid and prinophyllum (Small) Millais (King 1977; morphologically at the Gregory Bald site, consistent with R. calendulaceum (W4),
192 FALL 2012 Table 1 continued.
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though lacking a blotch (Table 1, Fig. ploid plants (W5, W6, W7, W9, W11, triploid Pentanthera azaleas have been re- 2, Google Earth Map, http://www.ces. and W12) were fairly typical of R. calendu- ported by Jones et al. (2007). We have also ncsu.edu/fletcher/mcilab/publications/ laceum, yet most lacked a blotch, and W5, documented three triploids (e.g., R. ‘Cha- pentantheramap.kmx). !ree other plants W6, and W7 had eglandular pedicel hairs. meleon’) in hybrid zones of R. calendula- from Wayah were diploid and fairly !e diploid plant W8 had a yellow "ower ceum and R. periclymenoides found along typical of R. arborescens (W1, W2, and with red pistil and stamens and an unusu- Lake Keowee, S.C. (Towe 2004; unpub- W3). However, all three had wavy petal ally long corolla (50.52 mm), suggesting lished data). margins, W2 and W3 had yellow blotches, hybridization between R. arborescens and Introgression between diploid and and W3 lacked any fragrance, possibly either R. calendulaceum or R. cumberland- tetraploid plants can occur in various ways. indicating some introgression from either ense. !e triploid specimen, W10, had an Diploid plants may produce an unreduced R. calendulaceum or R. cumberlandense (if orange-pink "ower with an orange blotch gamete (2n = 2x = 26) that combines with it has existed in this area). and red pistil and stamens, consistent with a reduced gamete from a tetraploid (1n = an interploid hybrid between R. calendula- 2x = 26) to produce a tetraploid with bal- Wine Spring Bald (W5-W12) ceum and R. arborescens. Although triploid anced genetic contributions from each Of the plants surveyed at Wine plants do appear to be rare, Li (1957) also parent (Bretagnolle and !ompson 1995; -spring Bald, six were tetraploid, one was reported apparent triploids (2n ~ 30-40) Galletta and Ballington 1996; Ramsey and diploid and one was triploid. !e tetra- on Wayah Bald. Other naturally occurring Schemske 1998). B019 and B056 from
JOURNAL AMERICAN RHODODENDRON SOCIETY 193
?C Gregory Bald could have arisen in such a be rare, once they do occur, they can pro- periods of melting-pot sympatry ostensibly manner with an unreduced gamete from vide a gateway for introgression. In studies led to rehybridization and reinforcement of R. arborescens or an R. arborescens hybrid of hybrid Iris zones, hundreds of advanced reproductive compatibility that e$ectively crossed with R. calendulaceum to produce hybrid genotypes were detected (e.g., F2, compromised divergence (Levin et al. a fragrant tetraploid in either F1 or subse- F3, BC2 BC3) but no F1s were found (Ar- 1996; Rieseberg 1997). !ese interludes quent generations. Unreduced gametes are nold 2006). !is observation suggests that would also have generated novelty and relatively common in Rhododendron spp. F1 hybrids may represent the most restric- diversity and spread successful alleles that (Widrlechner et al. 1982; Widrlechner tive bottleneck for genetic exchange, but allowed for enhanced #tness in new and and Pellett 1983). In many cases, hybrids once they are formed, they serve as bridges varied habitats during expansion phases have been found to produce unreduced for further exchange. Rare events of inter- (Arnold 1997; 2006; Riseberg and Carney gametes at higher frequencies than paren- speci#c/interploid hybridization can thus 1998). Although this ongoing genetic tal species (Ramsey and Schemske 1998). be extremely important and have substan- mixing appears to have contributed to a Alternatively, triploids can also serve as tial evolutionary impacts (Mallet 2005). poorly di$erentiated web of species, this genetic and reproductive bridges. Triploid Infrequent triploids, like the one found at retained capacity for hybridization most Rhododendron spp., though somewhat Wine Spring Bald, may play a key role in likely a$orded evolutionary and adaptive infertile, can produce unreduced gametes introgressive hybridization between R. ca- bene#ts for the common Pentanthera at relatively high (>5%) frequencies lendulaceum and R. arborescens there. gene pool, a mechanism ideally suited to (Jones and Ranney 2009), often at higher Habitats can change dramatically a rapidly changing climate. frequencies than diploids (Dweikat and over both the short- and long-term. Accumulated evidence suggests that Lyrene 1988; Veilleux 1985). !e union of Natural events like succession, #re, and hybridization among the Pentanthera an unreduced gamete from a triploid (2n climate shifts lead to ever-changing azaleas is more than a taxonomic muddle- = 3x = 39) with a normal reduced gamete distribution patterns and new species ment and botanical curiosity; rather, it from a diploid (1n = 1x = 26) o$ers an associations. As recently as 18,000 years re"ects a pervasive and ongoing evolution- alternative pathway, sometimes called a ago, during the last glacial maximum, the ary process that has provided these plants triploid bridge, for diploids to introgress/ Southern Appalachian region consisted with a successful adaptive outcome over transform into tetraploids (Ramsey and of boreal and tundra environments recent millennia. !e process has been less Schemske, 1998). (Delcourt and Delcourt 1975; Shafer of a breakaway divergence of individual Introgression from triploids and tet- 1988). Reconstructed paleovegetation lineages and more of an ongoing, inter- raploids into diploid populations is less maps for the Holocene epoch illustrate twined, evolutionary exchange. common, but can occur. In some cases, the continuous nature of climate-induced tetraploids can produce diploid o$spring changes in vegetation biomes in this region, Conclusion through the parthenogenetic development often without modern analogs (Overpeck Results of this study provide new of reduced gametes, a form of apomixis et al. 1992). On an even broader scale, evidence of both interspeci#c and sometimes referred to as polyhaploidy (de throughout the Pleistocene epoch, the interploid hybridization among species Wet 1971; Shoemaker 1986). !e pro- "ora of Southeastern North America of Pentanthera azaleas. Despite sympatric duction of 1x or ~1x gametes by triploids were subjected to multiple glacial cycles populations of diploids and tetraploids and tetraploids through complement frac- (as many as 20 over a period of 2 million with overlapping blooms, triploid hybrids tionation, asynchronous meiotic rhythms, years) (Davis 1983; Hays et al. 1969). were rare at these sites. However, rare chromosome elimination, random chro- Temperate species, including Pentanthera triploids can provide bridges for genetic mosome assortment, fertile aneuploids, azaleas, probably survived these glacial exchange between diploids and tetraploids. and other unusual meiotic pathways may periods on higher terrain at refugial Other reproductive mechanisms, also allow for introgression from higher to sites throughout southern Tennessee, particularly unreduced gametes, may lower ploidy levels (Avers 1954; Burton southeastern and northern Louisiana, also allow for direct genetic exchange and Husband 2001; Husband 2004; Lim southwestern Mississippi, southwestern between ploidy levels. Furthermore, the et al. 2005; Norrmann and Quarin 1987; Georgia, Alabama, and northern Florida substantial reproductive compatibility Peckert and Chrtek Jun. 2006; Ramsey (Davis 1976; Delcourt 1980; Parks et among species, ongoing hybridization, and Schemske 1998; Risso-Pascotto et al., al. 1994; Watts and Stuiver 1980). Over and the phytogeograpic history suggest 2004; Smith-White 1948; Widrlechner et time, surviving lineages appeared to have that the Pentanthera azalea species in the al. 1984). undergone cycles of southern migrations Southeastern United States continue to Although instances of interspeci#c followed by re-radiation that contributed evolve with a high level of genetic exchange and interploid hybridization may initially to a reticulate pattern of evolution. !ese through a successful evolutionary web.
194 FALL 2012 Figure 1. Topography map of Gregory Bald showing the location of Figure 2. Topography map of Wayah and Wine Spring Balds showing sampled plants. Triangle symbols indicate tetraploids and squares the location of sampled plants. Triangle symbols indicate tetraploids indicate- !diploids. #!* ! # #* $ ( % % $ ! ! %$-# $* $ -and squares ! #!* ! indicate * !# diploids with the exception$$ ( % of %W10 $ ! ! %$-being a triploid. %%%#! $ $"&#$ %! $- # $* $ %%%#! $ $"&#$ %! $(%%)!% 10 %#! -
Glen Jamieson 7/17/12 9:27 AM Deleted: ( %$- $#' % ## $+ $ %%#+# #+ ' ! '!!"$!! # ' $ !!$#"#$ # ! # ' !#$ #$! ##$+ * %$+ # *+ #$ #$% #$$$% (% % +# # + !#$$ + & % $ ! $-$$$% (% # ++## %!"#'+ "+ '" ! "" ! !#$ #$! # ++## !! # % * $ $ #% * !!#%- %!"#' ! # $#"#! #!+' !#* !# # Glen Jamieson 7/17/12 9:27 AM Formatted: Font:Italic !#$ #$! ) $ # !#$ #$! ! " "! &# " #!)/00 "!!+)
!#! ## %!"#') " %!) .2103+ !!* "!- "$+$) '* Glen Jamieson 7/17/12 9:28 AM Deleted: # ,! '- "$+$+ !"!#! ! "" ! " "" #!#$!+ Glen Jamieson 7/17/12 9:28 AM Deleted: * !#! !- &+ #- # ($#%#$$$% %$&!# % % &% Glen Jamieson 7/17/12 9:28 AM Deleted: 24 %&% # ! ( "+2003+ ##$$- #% % (. $ &%- %- Glen Jamieson 7/17/12 9:28 AM Deleted: , 25 Acknowledgements Avers, C.J. 1954. Chromosome behavior Burton, T.L. and B.C. Husband. 2001. !anks are given to Jared Barnes, in fertile triploid aster hybrids. Genet. Fecundity and o$spring ploidy in Jason Lattier, Irene Palmer, Kevin Parris, 39(1): 117-126. matings among diploid, triploid and Kelly Oates, Amira Ranney, and Chris Barrier, M., B.G. Baldwin, R.H. tetraploid Chamerion angustifolium Wendehorst for assisting with collecting, Robichaux, and M.D. Purugganan. (Onagraceae): Consequences for recording, pressing, and mounting 1999. Interspeci#c hybrid ancestry tetraploid establishment. Hered. 87: samples. Assistance with mapping from of a plant adaptive radiation: 573-582. Nathan Lynch is also greatly appreciated. Allopolyploidy of the Hawaiian Comai, L. 2005. !e advantages and silversword alliance (Asteraceae) disadvantages of being polyploidy. Nat. Literature cited inferred from "oral homeotic gene Rev. Genet. 6: 836–846. Adams, K.L. and J.F. Wendel. 2005. duplications. Mol. Biol. Evolution. Darwin, C. 1859. On the origin of species Novel patterns of gene expression 16(8): 1105-1113. by means of natural selection, or the in polyploidy plants. Trends Genet. Bowers, C.G. 1968. Rhododendron preservation of favoured races in the 21(10): 539-543. and azaleas: !eir cultivation and struggle for life. John Murray, London. Arnold, M.L. 1997. Natural hybridization development. 2nd ed. Macmillan Co., Davis, M.B. 1976. Pleistocene biogeog- and evolution. Oxford Univ. Press, N.Y. New York. raphy of temperate deciduous forests, Arnold, M.L. 2006. Evolution through Bretagnolle, F. and J.D. !ompson. 1995. p. 13-26. In: R.C. West and W.G. genetic exchange. Oxford Univ. Press, Tamsley review no. 78. Gametes with Haag (eds.). Geoscience and man, ecol- N.Y. the somatic chromosome number: ogy of the Pleistocene, vol. XIII. School Arnold, M.L. and E.J. Larson. 2004. Mechanisms of their formation and of Geoscience, Louisiana State Univ., Evolution’s new look. Wilson Qrtly., role in the evolution of autopolyploid Baton Rouge. Autumn: 60-72. plants. New Phytol. 129: 1-22. Davis, M.B. 1983. Quaternary history
JOURNAL AMERICAN RHODODENDRON SOCIETY 195 of deciduous forests of eastern North tocene sediments of the equatorial Leach, D.G. 1958. !e re-creation of America and Europe. Ann. Missouri Paci#c: !eir paleomagnetic, biostrati- a species. J. Amer. Rhodododendron Bot. Garden 70: 550-563. graphic, and climatic record. Geol. Soc. Soc. 12(4),
196 FALL 2012 Mallet, J. 2005. Hybridization as an 28: 359-389. Wisconsin climate of northern Florida invasion of the genome. Trends Ecol. Rieseberg, L.H. and S.E. Carney. 1998. and the origin of species-rich deciduous Evolution 20: 229-237. Plant hybridization. New Phytologist forest. Sci. 210: 325-327. Millais, J.G. 1924. Rhododendrons and the 140: 599-624. Willingham, F.F., Jr. 1976. Variation and various hybrids. Longman, Green, and Risso-Pascotto, C., M.S. Pagliarini, C. phenological forms in Rhododendron Co., London. Borges do Valle, and L. Jank. 2004. calendulaceum. Castanea 41: 215-223. Norrmann, G.A. and C.L. Quarín. 1987. Asynchronous meiotic rhythm as Widrlechner, M.P. and H.M. Pellett. Permanent odd polyploidy in a grass the cause of selective chromosome 1983. Notes on the pollen of deciduous (Andropogon ternatus). Genome 29: elimination in an interspeci#c azalea cultivars. J. Amer. Rhododendron 340-344. Brachiaria hybrid. Plant Cell Rep. 22: Soc. 37: 210-212. Osborn, T.C., J.C. Pires, J.A. Birchler, 945-950. Widrlechner, M., H. Pellett, and P. Ascher. D.L. Auger, Z.J. Chen, H. Lee, L. Sax, K. 1930. Chromosome stability in 1982. Unreduced gametes in azalea Comai, A. Madlung, R.W. Doerge, V. the genus Rhododendron. Amer. J. Bot. hybrids: A possible breeding method Colot, and R.A. Martienssen. 2003. 17(4): 247-251. for using promising azaleas of low Understanding mechanisms of novel Shafer, D.S. 1988. Late quaternary fertility. J. Amer. Rhodododendron Soc. gene expression in polyploids. Trends landscape evolution at Flat Laurel 36: 98-100. Genet. 19(3): 141-147. Gap, Blue Ridge Mountains, North Widrlechner, M.P., H.M. Pellett, and Overpeck, J.T., R.S. Webb, and T.Webb Carolina. Quat. Res. 30: 7-11. P.D. Ascher. 1984. Observations of III. 1992. Mapping eastern North Skinner, H.T. 1955. In search of native phylogenetic relationships between American vegetation change of the azaleas. Morris Arboretum Bull. 6:1-10, Rhododendron calendulaceum and past 18 ka: No-analogs and the future. 15-22. R. bakeri. J. Rhododendron Species Geology 20: 1071-1074. Skinner, H.T. 1961. Classi#cation of the Foundation 1: 49-63. Parks, C.R., J.F. Wendel, M.M. Sewell, native American azalea. p. 81-86. In: and Y.-L. Qiu. 1994. !e signi#cance Proc. Intern. Rhododendron Conf., of allozyme variation and introgression Amer. Rhododendron Soc., 11-14 in the Liriodendron tulipifera complex May, Portland, Ore. (Magnoliaceae). Am. J. Bot. 81(7): Smith-White, S. 1955. !e life history Leucopogon 878-889. and genetic system of Kimberly Shearer is currently an under- juniperinus Heredity Peckert, T. and J. Chrtek Jun. 2006. . 9: 79-91. graduate student majoring in Plant Mating interactions between coexisting Soltis, D.E., P.S. Soltis, L.H. Rieseberg. Biology and Horticultural Science at N.C. diploid, triploid and tetraploid 1993. Molecular data and the dynamic State University. !omas G. Ranney is a Critical Rev. Plant cytotypes of Hieracium echioides nature of polyploidy. Professor of Horticultural Science at N.C. Sci. (Asteraceae). Folia Geobotanica 41(3): 12: 243-273. State University and a member of the 323-334. Soltis, D.E., P.S. Soltis, and J.A. Tate. Southeastern Chapter. !ey can be reached Ramsey J. and D.W. Schemske. 1998. 2003. Advances in the study of at: Department of Horticultural Science, New Pathways, mechanisms, and rates polyploidy since plant speciation. Mountain Horticultural Crops Research and Phytol of polyploid formation in "owering . 161: 173-191. Extension Center, 455 Research Dr., North plants. Annu. Rev. Ecol. Syst. 29: 467- Soltis D.E., V.A. Albert, J. Leebens-Mack, Carolina State University, Mills River, NC 501. C.D. Bell, A.H. Paterson, C. Zheng, 28759. Kim Shearer: kmsheare@ncsu. Rehder, A. 1921. !e azaleas of North D. Sanko$, C.W. dePamphilis, P.K. edu, Tom Ranney: [email protected]. America, p. 107-196. In: E.H. Wilson Wall, and P.S. Soltis. 2009. Polyploidy Ron Miller is a retired professor of English Amer. and A. Rehder (eds.). A monograph of and angiosperm diversi#cation. Renaissance literature. He can be reached azaleas. J. Bot. 96(1): 336-48. !e Arnold Arboretum Pub. at [email protected]. Clarence Towe is a American Azaleas #9, University Press, Cambridge, Mass. Towe, L.C. 2004. . retired assistant superintendent of education Diploid pollen Timber Press, Portland, Ore. Shoemaker, B.L. 1986. and the author of American Azaleas, 2004, viability characteristics and diploid- Veilleux, R. 1985. Diploid and polyploid Timber Press. He can be reached at lctowe@ tetraploid crossability in Vaccinium. gametes in crop plants: Mechanisms bellsouth.net. M.S. !esis, N. C. State University, of formation and utilization in plant Raleigh, NC. breeding. Plant Breeding Rev. 3: 253- Rieseberg, L.H. 1997. Hybrid origins of 288. plant species. Ann. Rev. of Ecol. and Sys. Watts, W.A. and M. Stuiver. 1980. Late
JOURNAL AMERICAN RHODODENDRON SOCIETY 197