G. DIERINGER, L. CABRERA R., 2000 31 A comparison of size and sexual expression in populations of macrospathum Benth. and A. dracontium (L.) Schott ()

G. Dieringer and L Cabrera R. Department of Biological Sciences University of Texas at Brownsville Brownsville, TX 78520, USA

ABSTRACT KEYWORDS

The sexual expression or breeding sys­ Arisaema macrospathum, Araceae, sex­ tem for tropical populations of Arisaema ual expression, cloud forest, Mexico. macrospathum was compared with that of its temperate counterpart, A. dracontium. INTRODUCTION Although A. macrospathum populations Arisaema macrospathum Benth., A. are morphologically very similar to A. dra­ dracontium (L.) Schott and A. triphyllum contium, they differ markedly in sexual (L.) Schott are the three Arisaema species expression. Temperate populations of A. found in North America (Correll & John­ dracontium are andromonoecious, while ston, 1979; Gleason & Cronquist, 1991; those of A. macrospathum occurring in McVaugh, 1993). These perennial species cloud forest were found to bear only uni­ are distributed mostly in temperate regions sexual flowers, staminate or pistillate with from southern Canada (Ontario and Que­ monoecious individuals completely lack­ bec; Yang et aI., 1999) through the eastern ing. Male produced a mean of 42.7 United States and south into Mexico, par­ flowers/spadix while females produced a ticularly along the montane east coast mean of 60.9. For males, number of flow­ (Johnston et ai., 1989: McVaugh, 1993). Ar­ ers/spadix showed a significant, positive isaema dracontium and A. triphyllum relationship with size as measured have been the subject of a number of eco­ by basal stern diameter, but no relation­ logical and evolutionary studies (Bierzy­ ship was detected for females. Non-flow­ chudek, 1982, Lovett-Doust & Cavers, ering and male plants did not differ in size, 1982, Clay, 1993, Yang et ai., 1999) and but were significantly smaller than female both have long been known to change sex plants as is typical for the congeneric A. between seasons (Schaffner, 1922; Bier­ triphyllum and probably switch gender zychudek, 1984). Sexual expression of in­ under similar circumstances also. Consid­ dividual plants is usually related to plant ering the entire geographic range, A. dra­ size and environmental factors affecting contium exhibits a gradual decline in plant plant size such as nutrition, leaf area, and size, number of flowers/spadix for males, size. and an increase in the expression of fe­ Typically, populations of A. dracontium male plants along a north-south gradient. are composed of male plants bearing only This gradual decline in size would appear staminate flowers ormonoecious plants to extend into populations of A. macros­ bearing staminate and pistillate flowers pathum and suggests a close evolutionary (Schaffner, 1922; Lovett-Doust & Cavers, relationship. Differences in sexual expres­ 1982; Clay, 1993; Boles, 1996); the breed­ sion between the two species may be re­ ing system is thus said to be andromon­ lated to geographical isolation and the oecious. On rare occasions, strictly female presence or absence of competing con­ plants bearing only pistillate flowers have generiC species. been noted. In studies by Schaffner (1922) 32 AROIDEANA, Vol. 23 involving 872 plants, and Lovett-Doust & and spadices were present although Cavers (982) involving 964 plants, no fe­ spathes were dry and papery. The popu­ male individuals were ever observed. lation size and density were estimated by However, in a study by Clay (993) in­ marking off the area containing plants volving 338 individuals a single female in­ with a meter tape and counting all individ­ dividual was recorded. It may be noted uals. In addition whether plants had flow­ that this last study was conducted around ered or not and sexuality of flowering Baton Rouge, Louisiana which is in the plants was recorded. All plants within the southern portion of the species' distribu­ population plus additional plants encoun­ tion. Populations in Mexico are considered tered nearby were used to record data on a distinct species, A. macrospathum (Cor­ number of flowers produced for each rell & Johnston, 1979; Johnston et al., plant and basal stem diameter as a mea­ 1989; McVaugh, 1993). McVaugh has de­ sure of plant size (Clay, 1993). A total of scribed A. macrospathum as being dioe­ 47 past-flowering plants were sampled, 33 cious since all individuals produce only from the study population and 14 from the one flower type, staminate or pistillate. surrounding area. Spadices of all 47 plants Over the past few years, we have been were fixed in alcohol and returned to the studying the pollination ecology and floral lab to count the number of flowers under thermogenicity of Magnolia tamaulipana a dissecting scope. Basal stem diameter found in cloud forest of the Biosphere Re­ was measured using a vernier caliper ac­ serve El CieJo, Tamaulipas, Mexico (Dier­ curate to 0.01 mm. inger et al., 1999). It was here that we be­ gan a preliminary study on reproductive RESULTS biology of A. marcrospathum in cloud for­ est. Data collection included population The population area measured 50 X 25 size, population density, number of flower meters and contained 147 plants for a den­ and non-flowering individuals, number sity of 0.12 plants/m2. Of the 147 plants, and sex of flowers produced per individ­ 33 (22.4%) were flowering and of these 33, ual, and plant size. 2406.3%) were males and 9 06.1%) were females. Only plants with unisexual flow­ MATERIALS AND METIIODS ers, staminate or pistillate, were observed. No monoecious individuals were ever not­ The El Cielo Reserve lies just south of ed while non-flowering and male plants the Tropic of Cancer within the Sierra Ma­ did not differ in size, female plants were dre Oriental along the eastern coast of Significantly larger than either non-flow­ northern Mexico (23°12'-23°03'N and ering or male plants (Table 1). Non-flow­ 99°18'-99°07'W; for a detailed description ering plants ranged in size from 2.8-7.8 of the area, see Puig et al., 1987). Cloud mm in basal stem diameter, male plants forest (bosque mes6filo de montana; Rze­ ranged from 2.4-7.3 mm, and female dowski, 1978) is a moist, montane vege­ plants ranged from 5.0-15.0 mm. tation type and in eastern Mexico typically Male plants produced a mean of 42.7 contains both temperate (e.g. Quercus, flowers/spadix (SE = 1.9, range 30-75) CaJpinus, Liquidambar, Platanus, Pinus) and female plants a mean of 60.9 flowers/ and tropical (e.g. Engelhardtia, Clethra, spadix (SE = 3.0, range 34-82; Table 2). Dendropanax, PodocaJpus) elements. At The relationship between plant size and El Cielo, cloud forest vegetation typically flower production was examined using occurs between 800-1,600 m. The study correlation. Number of flowers/spadix population was located along side the showed no significant relationship with road running between the ejidos San Jose plant size or basal stem diameter for fe­ and La Gloria at approximately 1,600 m males (r = -0.09, N = 20, P = 0.72; Pear­ elevation. On May 20, 1998, plants in the son correlation) (Fig. 1) while male plants study area were past-flowering; all spathes showed a Significant, positive relationship G. DIERINGER, 1. CABRERA R., 2000 33

Table 1. One-way ANOVA comparison of mean basal stem diameter (mm) between non-flowering, male, and female sexual morphs of Arisaema mac- rospatbum at the Biosphere Reserve EI Cielo in 1998. Mean Sexual basal stem morph Np1ants diameter Range SE F P Non-flowering 25 4.43" 2.8-7.8 0.24 Males 27 5.09" 2.4-7.1 0.23 20.74 0.0001 Females 20 7.41h 5.0-15.0 0.52 Means sharing the same letter are not significantly different (P > 0.05, Scheffe's multiple com- parison test).

(r = 0.39, N = 26, P < 0.05; Pearson cor­ of changing sex. Our above observations relation) (Fig. 1). would then suggest they represent a situ­ ation similar to A. triphyllum where most DISCUSSION individuals produce only unisexual flow­ The study population, as well as all ers, but are capable of switching gender plants of Arisaema macrospathum ob­ and that populations of A. macrospathum served at the Biosphere El Cielo, produced should not be categorized as dioecious. only unisexual flowers. No monoecious Studies by Clay (1993) indicate that individuals were observed. This is exactly some American populations of A. dracon­ as described for A. macrospathum by tium also have the potential to produce McVaugh (1993) although he categorized female plants. In Clay's study of 339 the species as dioecious. plants, one female plant was observed Previous studies by Schaffner (1922) measuring 12 mm in diameter and pro­ have demonstrated that both A. dracon­ ducing 77 flowers, values similar to those tium and A. triphyllum are capable of at El Cielo. In addition, Clay reported that changing sex. We presume that popula­ plants growing in a botanical garden in tions of A. macrospathum are also capable North Carolina may eventually produce

Table 2. A comparison of mean number of flowers per plant for male, mon- oecious, and female sexual morphs in populations of Arisaema dracontium andA. macrospatbum along a north-to-south latitudinal gradient. Parentheses contain sample size. Male Female plants Monoecious plants plants

Population #0 #0 #~ #~ Reference Arisaema dracontium OntariO, Canada 133 80 110 - Lovett- Doust (35) (35) & Cavers, 1982 Louisiana, USA 120.2 78.8 89.2 77 Clay, 1993 (171) (167) (1) Arisaema macrospathum Tamaulipas, Mexico 42.7 - - 60.9 this study (27) (20) 34 AROIDEANA, Vol. 23

100 A dracontium, male plants of A. macrospa­ U) 80 tbum produced fewer flowers/spadix but 0:: W were similar in size (Tables 1, 2 and 3). ~ 60 0 ConSidering both species, a north-to-south ...J • u. 40 latitudinal gradient can be discerned for ci z 20 size of plants, flower number of males and expression of individuals bearing only pis­ 0 tillate flowers. In southern populations, 0 5 10 15 20 plants are smaller in stem diameter, males STEM DIAMETER (MM) produce fewer flowers/spadix, and the number of pure pistillate flowered individ­ uals increases. 80 .8 The relationship between number of U) flowers/spadix and plant size do not fit 0:: 60 W • any models for the evolution of female ~ 0 40 ·6._ .. plants from an originally andromonoe­ ...J •• • •• u. ••• • • cious population (Clay, 1993). Clay has zci 20 suggested that, in temperate areas, co-oc­ currence of A. dracontium with A. tri­ 0 pby/lum and pollinator sharing between 0 2 4 6 8 the two species may select for andromon­ STEM DIAMETER (MM) oecy to reduce competition for pollinators. To our knowledge, Arisaema tripbyllum Fig. 1. Scatter-plot of flower number per does not occur in Mexico. From this per­ plant versus basal stem diameter (mm) for spective, unisexual plants may be selec­ sexual morphs of Arisaema macrospa­ tively favored over andromonoecious tbum at the Biosphere Reserve EI Cielo, plants in the absence of pollinator com­ Tamaulipas, Mexico in 1998. A. Female petition, andromonoecy never having ex­ plants. B. Male plants. isted in the Mexican populations of A. ma­ crospatbum. only pistillate flowers. These data clearly The north-to-south gradient in flower indicate that populations of A. dracon­ number, plant size, and expression of fe­ tium, although andrdmonoecious in the male individuals observed for A. dracon­ field, possess the capability of producing tium extends directly into populations of only female individuals given the right en­ A. macrospatbum in northern Mexico and vironmental conditions. suggests these two species are derived Compared to previous studies on A. from a common ancestral population. Ar-

Table 3. Mean basal stem diameter (mm) for male, monoecious, and female sexual morphs in populations ofArlsaema draconUum along a north-to-south latitudinal gradient. Parentheses contain standard errors. Monoecious Population Male plants plants Female plants Reference Ontario-1 8.9 (0.1) 15.0 (0.4) - Boles, 1996 Ohio-1 9.0 (0.1) 14.1 (0.2) - Boles, 1996 Indiana-4 8.1 (0.7) 12.4 (0.4) - Boles, 1996 -l - 9.8 (0.4) - Boles, 1996 Louisiana, USA 6.3 (0.1) 10.1 (0.2) 12.0 Clay, 1993 Louisiana-l - 8.9 (0.5) - Boles, 1996 G. DIERINGER, L. CABRERA R., 2000 35 isaema dracontium and A. macrospathum 1881 pp. The University of Texas at would therefore be sister species. We Dallas, TX. speculate that the most likely mechanism Dieringer, G., 1. Cabrera R., M. Lara, 1. for the splitting of the original population Loya & P. Reyes-CastiIlo. 1999. Beetle into two species to be geographic isolation pollination and floral thermogenicity across the arid Rio Grande valley and into in Magnolia tamaulipana (Magnoli­ the mountains of Mexico and consequent aceae). Inter.]. PI. Sciences 160:64- reproductive isolation. The lack of com­ 71. peting congeneric species in Mexico se­ Gleason, H. A. & A. Cronquist. 1991. Man­ lected for plants bearing unisexual flowers ual of Vascular Plants of Northeast­ and therefore added to the genetic diver­ ern United States and Adjacent Can­ gence between the two groups. ada. 2nd ed., 910 pp. The New York Botanical Garden, NY. ACKNOWLEDGMENTS Johnston, M. C., K. Nixon, G. L. Nesom & M. Martinez. 1989. Listado de plantas The study was funded, in part, by the vasculares conocidas en la sierra de Department of Biological Sciences, West­ Guatemala, G6mez Fanas, Tamauli­ ern Illinois University, Macomb, IL 61455. pas, Mexico. Biotam 2:21-33. We thank Tom Croat for assistance in ver­ Lovett-Doust, ]. & P. B. Cavers. 1982. Re­ ifying specimen identifications. We also source allocation and gender in the thank Peter Boyce and two anonymous re­ green dragon Arisaema dracontium viewers for valuable comments on earlier (Araceae). Amer. Midi. Naturalist 108: drafts of the manuscript. 144-148. McVaugh, R. 1993. Flora Novo-Galiciana. UTERATURE CITED A descriptive account of the vascular Bierzychudek, P. 1982. The demography plants of western Mexico, vol 13. The of jack-in-the-pulpit, a forest peren­ University of Michigan Herbarium, nial that changes sex. Ecol. Monogr. Ann Arbor, MI. 52:335-351. Puig, H., R. Bracho & V.]. Sosa. 1987. El ---. 1984. Assessing "optimal" life his­ bosque mes6filo de montana: com­ tories in a fluctuating environment: posici6n flonstica y estructura. pp. the evolution of sex-changing by 55-79 in H. Puig, R. Bracho, Eds. EI jack-in-the-pulpit. A mer. Naturalist Bosque Mes6.filo de Montana de Ta­ 123:829-840. maulipas. Instituto de Ecologfa, A.c., Boles, R. 1996. Genetic structure and sex­ Mexico, D.F. ual reproductive potential of the pe­ Rzedowski,]. 1978. Vegetaci6n de Mexico, rennial aroid Arisaema dracontium, 432 pp. Editorial Limusa, Mexico, D.F. 140 pp. Master's TheSis, University of Schaffner,]. H. 1922. Control of the sexual Windsor, Ontario, Canada. state in and Ar­ Clay, K. 1993. Size-dependent gender isaema dracontium. Amer. ]. Bot. 9: change in green dragon (Arisaema 72-78. dracontium; Araceae). Amer. ]. Bot. Yang,].,]. Lovett-Doust, & 1. Lovett-Doust. 80:769-777. 1999. Seed germination patterns in Correll, D. S. & M. C.Johnston. 1979. Man­ green dragon (Arisaema dracontium, ual of the vascular plants of Texas, Araceae). Amer.]. Bot. 86: 1160-1167.