Torrey Botanical Society

Self-Fertility in Canada Yew ( canadensis Marsh.) Author(s): Taber D. Allison Reviewed work(s): Source: Bulletin of the Torrey Botanical Club, Vol. 120, No. 2 (Apr. - Jun., 1993), pp. 115-120 Published by: Torrey Botanical Society Stable URL: http://www.jstor.org/stable/2996940 . Accessed: 04/04/2012 14:40

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Torrey Botanical Society is collaborating with JSTOR to digitize, preserve and extend access to Bulletin of the Torrey Botanical Club.

http://www.jstor.org Bulletinof the TorreyBotanical Club 120(2), 1993, pp. 115-120

Self-fertilityin Canada yew (Taxus canadensisMarsh.)' Taber D. Allison2 Departmentof Biology, Ohio StateUniversity, Marion, OH 43302

ABSTRACT ALLISON, T. D. (Departmentof Plant Biology, The Ohio State University, Marion, OH 43302). Self-fertilizationin Canada yew(Taxus canadensisMarsh.). Bull. Torrey Bot. Club 120: 115- 120. 1993.-Self-fertilitywas examinedin a Canada yew(Taxus canadensis)population in southeasternMinnesota. Bagged yew branches and unbaggedcontrol branches showed no sig- nificantdifferences in production,proportion of pollinatedseeds that matured, or seed masses.Greater than 99% of male strobiliwere located on currentgrowth, but a significant proportionof female strobili (> 33%) werelocated on growthtwo years or older.There was no consistenttrend in locationof these strobilus types on higheror lowerbranches. Results from otherstudies examining consequences of monoecyin Canada yewsupport the conclusions of thisstudy that Canada yewis self-fertileand thatself-pollination enhances seed production whenpollen is limiting. Keywords: Taxuscanadensis, self-fertilization, plant reproductive biology, pollination biology.

The effectsof self-pollinationin gymnosperms sis) (Allison 1990a, 1990b). This species is the often are expressed by such post-zygoticphe- sole monoecious member of the genus Taxus nomena as early embryoabortion (indicated by (den Ouden and Boom 1982), and it is of interest reduced seed production or the presence of un- to ask whetherCanada yewis self-fertileor shows filledseeds), lower seed masses, and/orreduced any indication of inbreeding depression. Such seedlinggrowth. These phenomena characterize questions are relevant to understanding the inbreedingdepression-the reductionin fitness adaptive significanceof monoecyin an otherwise resultingfrom self-fertilization.Inbreeding de- dioecious genus. pression is apparentlycaused by the expression STUDY ORGANISM. Canada yew is an ever- of recessive lethal, or otherwisedeleterious, al- green indigenous to the mixed - leles in offspringproduced by self-fertilization hardwood forestsof northeasternUnited States (e.g., Dobzhansky et al. 1977). and southeasternCanada. The male and female Many coniferousspecies show some formof reproductivestructures (hereafter referred to as inbreedingdepression followingself-pollination strobili) typicallybegin to develop in the late (Snyder1968; Sorensen 1970, 1971; Dorman and summer in the axils of the year's current Squillace 1974; Ledig and Fryer 1974). Other growth(Dupler 1917). Maturationof the strobili ,however, show littleif any negativeef- and pollination of female strobilioccur the fol- fectof self-pollination(Fowler 1965a), and there lowingspring. Female strobiliare uniovulateand can be considerable variation in self-fertility when fertilizedproduce a singleripe seed thatis within species that otherwise show inbreeding surroundedby a red, fleshy,aril-like structure. depression (Hanover 1975). Seed ripeningbegins in late Julyor earlyAugust I have examined various aspects of the polli- and continuesfor 6-8 weeks into early fall. nation ecology of Canada yew (Taxus canaden- The growthform of Canada yew is diffuse; branches are typicallyascending and rarelyar- borescent(see Fig. 1). Vegetativepropagation of 'Denise Thiede and Steve Householderprovided Canada yew occurs when these ascending invaluableassistance in thefield. I thankGayle Muen- branchesare pressedto theground and take root. chowand an anonymousreviewer for their comments thatimproved this manuscript. This studywas funded Axillary branches offthe main stem axis may by grantsfrom the Dayton-WilkieFund of the Bell also take root, causing the plant to spread along Museumof NaturalHistory and theMinnesota Zoo- the forestfloor (Figs. 1 and 2). Connections be- logicalSociety. tween these rooted branches, or shoot systems, 2 Presentaddress: Harvard Forest, P.O. Box 68, Pe- tersham,MA 01366. can be tracedunder the litter layer, but eventually Receivedfor publication March 27, 1992, and in these connections rot. In high-densitypopula- revisedform August 20, 1992. tions of Canada yewgenetic relatedness of neigh- 115 116 BULLETIN OF THE TORREY BOTANICAL CLUB [VOL. 120

In Spring1984 and 1985at North Grey Cloud, I pairedterminal branches of neighboringyew 1-10 branch ,or shootsystems, (eight in 1984and nine in 1985)-that contained both female strobili and 4-2 D branch malestrobili. I had determinedthat these plants i m werenot connected belowground. I randomly as- signedone member of the pair of branches to the treatmentor control group. Treatment branches wereenclosed in parchment "Pollen-tector" bags (CarpenterPaper Company,Des Moines, IA 50302). Brancheswere bagged while male stro- bili wereimmature; no pollenhad been shedin thepopulation prior to bagging.Only a portion of thebranch could be coveredby thebag, and I labeledthe location of the end ofthe bag with a tag.The numberof ovules on treatmentand Fig. 1. Skeletonview of a hypotheticalshoot sys- controlbranches ranged from nine to 40. 1 bagged temof Canada yew.Primary and secondarybranches fouradditional branches, each on a separateplant, as describedin the textare indicated.For mapping strobilusposition, primary branches were numbered fromwhich all male strobilihad been removed consecutivelybeginning with the shoot tip. With the to determinewhether female strobili developed exceptionof the shoot tip, primary branches are equiv- in theabsence of pollination and to measurethe alentto axillarybranches. Dashed linescorrespond to extentof pollenleakage into the bags. belowgroundconnections to otheryew shoot systems. Bagswere removed one month after the pollen sheddingseason had ended.I countedthe num- boringplants, or more appropriately,shoot sys- berof developing on baggedand unbagged tem networksis difficultto determinebecause -of branches.Developing seeds wereeasily distin- this layeringhabit. guishedby their swollen, green appearance from non-developingseeds which were yellow and un- Methods. Fieldwork was performed at North changedin size. No mortalityof developing seeds GreyCloud Island,Minnesota (92?56'W longi- had occurredby thistime. I assumed,therefore, tude,44?46'N latitude) and at the Apostle Islands thatthis census was a measureof pollination National Lakeshore, Bayfield, Wisconsin success.I measuredseed production by counting (90?45'W longitude,46?50'N latitude). seedsor emptyreceptacles that indicated the re-

W,

Fig. 2. A denseCanada yew colony. 1993] ALLISON: SELF-FERTILITY IN CANADA YEW 117

Table 1. Mean and standarddeviation of (a) pollinationsuccess (pollinated ovules expressedas a proportion of total ovules), (b), seed set (ovules producingseeds expressed as a proportionof total ovules, and (c) mature seed productionexpressed as a proportionof pollinated ovules forbagged and unbagged Canada yew branches at Grey Cloud, MN 1984 and 1985. "n" refersto the number of branches. Analyses were performedon arcsin Y square root transformeddata. Results presentedin the table have been back-transformed.An asteriskby the mean or standard deviation indicates a significantdifference between treatmentsfor that parameterin that year. Pollination Proportionof success Seed set maturedseeds. n x SD SD x SD 1984 Bagged 8 0.78* .07 0.23 .02 0.39 .14* Unbagged 8 0.97* .03 0.24 .01 0.25 .01* 1985 Bagged 9 0.45** .15 0.14* .08 0.50 .31* Unbagged 9 1.00** .02 0.34* .02 0.35 .02*

*p < 0.05. ** p < 0.005. moval of a ripe aril and seed. All available ripe I was unable to germinate Taxus seeds during seeds on treatmentbranches were collected, re- the course of this study,so the relationshipbe- moved fromtheir arils, and air-driedfor at least tween-seed germination,seedling growth,and 6 months. Air-driedseeds were weighed to the selfingin Canada yew remainsunresolved. Seeds nearest milligram. of all members of the,genus Taxus have tena- I compared the effects.of treatmenton the pro- cious dormancy(Rudolf 1974; Melzack and Watts portion of ovules pollinated: (pollination suc- 1982). cess), proportionof total ovules.matured as seeds (seed set), and proportionof pollinated ovules SEED PRODUCTION.Unbagged brancheshad a matured as seeds (maturation level). Student's significantlygreater pollination success than t-tests(Sokal and Rohlf 1981) were performed bagged branches in both 1984 and 1985 (Table on arcsin \/Ytransformed proportions; Wilcox- 1). Baggedbranches also had a significantlylower on's matched pairs signed ranks test was used seed set in 1985, but there was no significant fortreatment comparisons when treatmentvar- differencein 1984. Maturation levels of bagged iances were unequal (Sokal and Rohlf 1981). brancheswere higherin both 1984 and 1985 but I recorded the branch.position of male and these differenceswere not significant. female strobilion seven plants in a Canada yew Pollination success in bagged branchesranged population on Stockton Island of the Apostle from33% to .100% in 1984 and from6% to 94% Islands. I designated' the terminal branch as in 1985, but the variances in pollination success branchone and thennumbered all distal primary did not differsignificantly by treatment.The low- branches (those coming offthe main stem axis) er pollination success of bagged branches could in sequence (i.e., branch two, branch three,etc.) reflectreduced pollen transferwithin the bags down to the base of the stem (Fig. 1). Both the ratherthan. differences in self-fertility.Variances branch numberand'the numberand sex of stro- in maturationlevels were significantlygreater in bili on the branchwere recorded.I compared the bagged branches(P < 0.05) in both.years(Table median branch (the:location of 50% of male or 1).-This could reflectdifferences in levels of in- female strobili)of each strobilustype using Wil- breedingdepression. among yews. coxon's two sample test(Sokal and Rohlf 198 1). In branchesbagged without male strobili,1.7% of the ovules began developing one month after Results and Discussion. Negative effectsof pollen shedding, but none of the developing self-fertilizationon plant fitnesscan be expressed. ovules matured,into seeds. Parthenocarpy,the in a varietyof ways:.(l) reduced seed production developmentof fullyformed, empty seeds in the resultingfrom the low. fertilizationof ovules,or absence of pollination,has been reportedin the low maturation percentages;.(2) reduced seed genus Taxus (Orr-Ewing 1957), but in this ex- masses and a greaterproportion of unfilled seeds; periment,I cannot rule out the possibilityof pol- and (3) low seed germinationand- seedling vigor. len leakage into the bags. However, the rarityof 118 BULLETIN OF THE TORREY BOTANICAL CLUB [VOL. 120

Table 2. Mean seed masses in mg (and 1 SE) of servedthis latter phenomenon in all Canada yew bagged and unbagged branches in Grey Cloud, MN populationsthat I have studiedalthough the per- Canada yew population for 1984 and 1985. The num- centage varies from population to population ber of seeds in each treatmentis indicated by n. (Allison unpublished data). The median branch 1984 1985 for male strobili was lower than the median n mass n mass branchfor female strobili in fourof seven plants. Bagged 11 30.94* (1.15) 31 26.13 (0.44) One plant had a lower median branchfor female Unbagged 25 28.09 (0.75) 21 26.96 (0.72) strobili,and in two plantsthere was no significant differencein the median branchfor the two stro- *p < 0.05. bilus types. In conifers,male strobiliare typicallylocated developing ovules in branches bagged without on lower branches (Smith 1981), and this has male strobiliindicates that the vast majorityof been suggestedas a means of reducingself-pol- ovules developing and maturing as seeds on lination in monoecious species (Fowler 1965b). branchesbagged with male strobiliwere self-pol- I did a simple statisticaltest that showed a slight linated. tendencyfor spatial separation of male and fe- male strobiliin Canada yew. More detailed anal- SEED MASSES. Seeds of bagged branches were ysis mighthave shown differencesin their dis- significantlyheavier in 1984; no differencein tribution,but the growthform of Canada yew seed masses was observedin 1985 (Table 2). Seed renderssuch differencessuperfluous. Canada yew masses have been shown to be positively cor- lacks strong apical dominance, and axillary relatedwith early seedling development in Pinus branches within a shoot system are often only strobus(Spurr 1944). Sorensenand Miles (1974), slightlylower than the terminal branch of the however, noted that 1st year growthof selfed main shoot axis (Fig. 1). In addition,Canada yew progenywas 18% to 21% lower than open-pol- has limited verticalgrowth; plants spread hori- linated progeny in Douglas-fir (Pseudotsuga zontally forminglarge dense clones. Ovules on menziesii) even though the selfed seeds of this one rametor shoot probablyreceive largequan- species showed negligibledifferences in seed mass titiesof pollen frommale strobilion othershoots compared to open-pollinated seeds. A lack of withinthe clone. inbreedingdepression in Taxus seedlings,there- Female strobiliwithin a plant may appear up fore, cannot necessarilybe inferredfrom seed to 12 days priorto the maturationof male stro- mass data alone. bili. Pollen dispersal may occur over a time pe- I collected102 ripeseeds fromtreatment plants riod as short as 24 hours (Allison personal ob- over two years. Fourteen of these seeds were servation).I do not know ifdifferences in timing empty,but the proportionof unfilledseeds did of maturationreduce self-pollination. not differbetween bagged and unbagged plants The results indicate that Canada yew is self- (X2 = 0.12). A high proportionof unfilledseeds fertile,and that thereare no obvious barriersto is a strongindication of high levels of self-pol- self-pollinationunder natural conditions. The linationin plants thatsuffer from inbreeding de- failureof this studyto detectinbreeding depres- pression (Sarvas 1962). Based on this criterion, sion can be explained in two ways. First, there treatmentyews show littleevidence of inbreed- is a high degree of selfingin naturalpopulations ing depression when compared to controlyews. of Canada yew. Ovules on unbagged branches STROBILUs LOCATION. Thirty-five percent of may have received considerableself-pollen from the primary branches in the seven measured the same branchor fromgenetically identical but Canada yew plants had both male and female separateshoot systems.High densityCanada yew strobili;the resthad eithermale or female stro- populations,such as at Grey Cloud, may consist bili. Typically,within a primarybranch system, of a low number of geneticallydistinct individ- eithermale or female strobilionly were located uals spreading clonally. Nearest neighbors do- on secondary branches, but a low percentage nating pollen to ovules on unbagged branches (< 5%) of secondarybranches had both male and could be members of the same genet. female strobili. Greater than 99% of the male If the above is the case, it would be more ap- strobiliwere located on currentgrowth (branch propriateto say thatI was comparingthe effects tissue produced the previous growingseason). of between-rametand within-rametpollination. More than 33% of the female strobiliwere lo- Many oftheseeds producedby unbagged branch- cated on growthtwo years or older. I have ob- es could also have resultedfrom self-nollination 19931 ALLISON: SELF-FERTILITY IN CANADA YEW 119 and both treatmentand controlgroups could be (and possibly the other members of the family showing the negative effectsof selfing.Female ) suggeststhat the most parsimonious plants that I created in 1986 at Grey Cloud had explanation is that monoecy is a derived trait. significantlylower pollination percentages,but Answersto questions concerningthe evolution- not seed production,than monoecious controls, ary origin (or maintenance) of monoecy within indicatingthat significantwithin-plant pollina- the genus Taxus as well as the geneticvariability tion occurseven in a highdensity yew population of Canada yew require furtherstudy. (Allison 1990b). To adequately test the effectof self fertilization,branches withoutmale strobili Literature Cited should be bagged and pollen from genetically differentyews applied to the ovules. ALLISON, T. D. 1990a. Pollen productionand plant In anotherexperiment, emasculated yews in a densityaffect seed productionin Taxus canadensis. Ecology 71: 516-522. low density yew population at the Apostle Is- 1990b. The influenceof deer browsingon the lands did not have highermaturation levels than reproductivebiology of Canada yew (Taxus can- monoecious yews (Allison 1990b). Both groups adensis Marsh) II. Pollen limitation: An indirect were naturallypollinated. Monoecious yews had effect.Oecologia 83: 530-534. DEN OUDEN, P. AND B. K. BooM. 1982. Manual of higher seed production than female yews, and cultivatedconifers. The Hague. 526 p. this differencepresumably was due to selfing. DOBZHANSKY, T., F. J.AYALA, G. L. STEBBINS AND J. Female yews in this population were all out- W. VALENTINE. 1977. Evolution. W. H. Freeman crossed.This resultsuggests that outcrossed seeds Co., San Francisco. 572 p. are not more likelyto maturethan selfed seeds. DoRMAN,K. W. AND A. E. SQUILLACE. 1974. Genetics of slash pine. USDA For. Serv. Res. Paper. WO- A second, and not mutually exclusive, hy- 20, 20 p. pothesisis thatCanada yew has low geneticvari- DUPLER, A. W. 1917. The gametophytesof Taxus abilityand has not accumulateda significantpro- canadensis Marsh. Bot. Gaz. 64: 115-136. portionof lethal recessive genes. Electrophoresis, FOWLER, D. P. 1965a. Effectsof inbreedingin red pine, Pinus resinosaAit. II. Pollination studies.Sil- forexample, detectedno geneticpolymorphisms vae Genet. 14: 12-23. in red pine (Pinus resinosa),which is highlyself- 1965b. Effectsof inbreedingin red pine, Pi- fertile(Fowler 1965a, 1965b; Fowler and Morris nus resinosa Ait. III. Factors affectingnatural self- 1977). Full evaluation of both hypothesesawait ing. Silvae Genet. 14: 37-46. furtherstudy. AND R. W. MoRRis. 1977. Genetic diversity in red pine: Evidence forlow genic heterozygosity. Finally,as mentionedpreviously, Canada yew Canad. J. For. Res. 7: 343-347. is the only monoecious member of the genus GHISELIN, M. T. 1969. The evolution of hermaph- Taxus, and giventhe self-fertilityof Canada yew, roditismamong animals. Quart. Rev. Biol. 44:189- it is of interestto speculate on the adaptive sig- 208. HANOVER, J.W. 1975. Geneticsof blue spruce.USDA nificanceof Canada yew's mating system.Ghi- For. Serv. Res. Pap. WO-28, 12 p. selin (1969) proposed that hermaphroditism(of LEDIG, F. T. AND J.H. FRYER. 1974. Geneticsof pitch which monoecy in plants is one type) may en- pine. USDA For. Serv. Res. Pap. WO-27, 14 p. hance reproductivesuccess when potentialmates MELZACK, R. N. AND D. WATTs. 1982. Variationsin are scarce. It is widely assumed that the effec- seed weight,germination, and seedling vigour in the yew ( L.) in England. J. Biogeogr. tiveness of wind-pollination is dependent on 9: 55-63. conditions that promote high concentrationsof ORR-EWING, A. L. 1957. Possible occurrenceof vi- airborne pollen (Whitehead 1983), and there is able unfertilizedseeds in Douglas Fir. For. Sci. 3: a strongrelationship between pollen production, 243-248. plant spacing, and pollination in Canada yew RUDOLF, P. 0. 1974. Taxus L. Yew, pp. 799-802. In USDA Seeds of woody plants in the United States. (Allison 1990a). Monoecy in Canada yew may, U.S. Dept. Agric.,Agric. Handb. 450. therefore,be selectively advantageous under SARvAs,R. 1962. Investigationson the floweringand conditions of reduced pollen availability (or the seed crop ofPinus silvestris.Comm. Inst. For. Fenn. low abundance of male "mates"). This hypoth- 53: 1-198. SMITH, C. C. 1981. The facultativeadjustment of sex esis is supportedby the greaterseed production ratio in lodgepole pine. Amer. Nat. 118: 297-305. of monoecious Canada yews relative to female SNYDER, E. B. 1968. Seed yield and nurseryperfor- yewswhen pollen limitsseed production(Allison mance of self-pollinatedslash pines. For. Sci. 14: 1990b). 68-74. SoKAL,R. R. AND F. J.ROHLF. 1981. Biometry:The It is not known whethermonoecy is the de- principleand practice of statisticsin biological re- rived or ancestral traitin the genus Taxus, but search,2nd ed. W. H. Freeman and Co., San Fran- the presenceof dioecy in all other Taxus species cisco. 859 p. 120 BULLETIN OF THE TORREY BOTANICAL CLUB [VOL. 120

SORENSEN, F. C. 1970. Self-fertilityof a centralOr- SpuRR,S. H. 1944. Effectof seed weightand seed egon source of Ponderosa pine. USDA For. Serv. originon the earlydevelopment of easternwhite Res. Paper PNW-109. 9 p. pine.J. Arnold Arbor. 25:467-480. . 1971. Estimate of self-fertilityin coastal WHITEHEAD, D. R. 1983. Wind pollination:Some Douglas-firfrom inbreeding studies. Silvae Genet. ecologicaland evolutionaryperspectives, pp. 97- 20: 115-120. 108.In L. Real [ed.],Pollination biology. Academic AND R. S. MILES. 1974. Self-pollinationeffects Press,New York. on douglas-firand ponderosa pine seeds and seed- lings. Silvae Genet. 23: 135-138.