Received 14 November 1995 Heredity 77 (1996) 388—395 Low genetic variation in Amentotaxus formosana Li revealed by isozyme analysis and random amplified polymorphic DNA markers CHIEH-TING WANGt, WEI-YOUNG WANG1, CHIA-HUA CHIANG, YA-NAN WANG & TSAN-PIAO LIN*t rSilv/culture Division, Taiwan Forestry Research Institute, 53 Nan-Hai Road, Taipei and Department of Forestry, National Ta/wan University, Taipei, Taiwan The objective of this research was to use random amplified polymorphic DNA (RAPD) and isozyme analysis to investigate genetic variation in narrowly distributed populations of Amentotaxus formosana Li. A total of 20 loci from 10 enzyme systems were analysed in 50 individual trees from each of the two natural populations. No isozyme variation was observed in the Tsatsayalai population. Phosphoglucose isomerase (Pgi-1) was the only polymorphic enzyme in the Tawu population, giving 5 per cent polymorphic loci with 0.008 expected heterozygosity. No genetic distance was found between these two populations using isozymes. Amentotaxus formosana demonstrated a high proportion of monomorphic RAPD fragments, about 79 per cent, for 20 arbitrary oligonucleotide primers. High similarity (0.994) was found between the Tawu and Tsatsayalai populations. RAPD markers provided further confirmation of the low levels of genetic variation in A. forinosana detected by isozyme analysis. The value of isozyme analysis was emphasized by the finding of the rare allele, Pgi-la, which was present only in the Tawu population. Based on the analysis of 110 individuals, representing 16 per cent of a native population, it was found that the younger tree category had a higher frequency of Pgi-la (0.125) than the older tree category (0.053), resulting in an expected heterozygosity of 0.250 and 0.105, respectively. It was inferred that the appearance of the Pgi-la allele could be the result of a mutation in the Tawu population and that selection is acting directly upon trees carrying this allele. Keywords: Amentotaxus formosana, genetic variation, isozyme, RAPD. Introduction covers about 225 ha. In 1988 the Council of Agri- formosana Li is endemic to Taiwan, where it is probably the most narrowly distributed natural reserves for A. formosana, because the popu- culture, Taiwan, designated these two sites as Amentotaxus lations were on the brink of annihilation. The gymnosperm species. It is highly endangered and has received worldwide attention (Farjon et al., 1993). It is a dioecious tree producing large and heavy seeds, climate and the components of the plant communities of these two habitats are very similar and belong to the warm temperate rain forest (Yang, 1994). The reversed J-type of the population structure, judged by the frequency distribution of breast height diameter (DBH) classes, indicated that A. formosana populations can grow continuously and but is poor in production. It grows in the broadleaved forests of Taitung Forest District (Tawu) and the Pingtung Forest District (Tsatsayalai), at an elevation ranging from 900 to 1300 m. The Tawu population has about 700 trees greater than 1 cm in stably under protection in these habitats (Yeh et a!., diameter and covers an area of about 86 ha. The Tsatsayalai population has about 800 trees and 1992; Yang, 1994). The primary objective of this research was to investigate genetic variation within and between * these two populations of A. forinosana, using Correspondence. 388 1996 The Genetical Society of Great Britain. LOW GENETIC VARIATION IN AMEN TO TA XUS FORMOSANA 389 isozyme and random amplified polymorphic DNA (RAPD) markers (Williams et al., 1990). Genetic studies based on isozyme data have major advantages over RAPD markers in that they are cheaper and easy to perform, but also give more information on genotypic relationships. Some evidence has suggested that allozyme variation may not be able to provide an accurate or complete measure of nucleotide variation in the genome (Wolff, 1991; Heun et al., 1994; Meijer et al., 1994). RAPD markers, on the other hand, may provide a less biased measure of genetic variability and a greater resolution of subtle genetic differences for inferring genetic structure. RAPD analysis has resulted in a more definitive grouping (Heun et a!., 1994; Maal3 & Klaas, 1995), even though RAPD polymorphism is poorly under- stood and believed to be based upon either sequence variation or mismatches in the primer binding sites. However, in this report we present some unique information from the isozyme analysis that it is not possible to obtain using RAPD markers. Materials and methods Sampling The locations of the two natural populations, Tawu and Tsatsayalai, are shown in Fig. 1. Fifty trees were sampled from each population for isozyme analysis, and 25 and 20 individuals from the Tawu and Tsatsayalai populations, respectively, for RAPD analysis. Fewer individuals were included in the RAPDs than the isozyme analysis because of practical constraints. Random sampling was applied to the trees; however, uniformity was not possible because some trees were 6-phosphogluconate dehydrogenase (6PGD, EC 1.1.1.43); phosphoglucose isomerase (PGI, EC 5.3.1.9); phosphoglucomutase (PGM, EC 5.4.2.2); shikimate dehydrogenase (SKDH, EC 1.1.1.25); superoxide dismutase (SOD, EC 1.15.1.1). Young leaf tissue and megagametophytes were ground with extraction buffer (Feret, 1971). Electrophoresis and staining followed the procedures described by Cheliak & Pitel (1984). Isozyme data analysis Allele frequencies were calculated for each locus and population. The following four measures were used to quantify genetic variation within a population: (1) the expected heterozygosity (Nei, 1975) at each locus was calculated as k I He=1 D2 i= 1 where P, is the frequency of the ith allele, summed over k alleles; (2) the mean number of heterozygous loci per individual was calculated (Nei, 1973); (3) the mean number of alleles per locus was calculated by averaging over all polymorphic and monomorphic loci; and (4) the effective number of alleles per locus (Ae; Crow & Kimura, 1970), was defined as Ae 1/P?. The number of alleles is maximized when the allele frequencies at any locus are equal. Both Wright's (1969) F-statistics and Nei's (1978) unbiased genetic identity (Ia) and genetic distance (D,,) were used to quantify the degree of differentiation among populations. The above calculations, with the exception of growing on a steep slope. Young leaf tissue was collected in April 1994 and stored at —20°C until required and additional seeds were collected in the effective number of alleles, were performed January 1995 for gametophyte analysis. Young leaf tissue of an additional 60 individuals in the Tawu population was collected in April 1995, and a total of 110 individuals, which varied from <5 cm to 47 DNA preparation cm DBH, were used to compare the genotype distribution and allele frequencies of Pgi-l. Isozyme electrophoresis methods Horizontal starch gel electrophoresis was used to examine ten enzyme systems, namely: esterase (EST, EC 3.1.1.1); fluorescent esterase (F-EST, EC 3.1.1.1); L-aspartate aminotransferase (AAT, EC 2.6.1.1); isocitrate dehydrogenase (IDH, EC 1.1.1. 42); malate dehydrogenase (MDH, EC 1.1.1.37); The Genetical Society of Great Britain, Heredity, 77, 3 88—395. using BIOSYS-1 (Swofford & Selander, 1989). Total cellular DNA was prepared from 0.8 g of young leaf material using a modified mini-CTAB method (Murray & Thompson, 1980). Leaves were frozen in liquid nitrogen, ground to a fine powder and suspended in 15 mL extraction buffer (50 mM Tris—HC1, pH 8.0, 350 mrvt sorbitol, 5 mivi sodium EDTA, 10 per cent polyethylene glycol 3350, 0.1 per cent bovine serum albumin, 0.1 per cent spermine, 0.1 per cent spermidine and 0.1 per cent 2-mercaptoethanol). The extraction was filtered through miracloth and centrifuged at 13 000 g for 15 mm in a Kontron H-401 centrifuge. The pellet was resuspen- ded in 350 L resuspension buffer (50 mi Tris— 390 C.-T. WANG ETAL. N 0 I 2 3KM Fig. 1 The natural reserves for Amentotaxus forinosana. A, Tawu; B, Tsatsayalai. HC1, pH 8.0, 25 mivi EDTA, 350 mrvi sorbitol and 0.1 per cent 2-mercaptoethanol) and the nuclei and organelles lysed by addition of 25 #L 20 per cent sarkosyl (N-lauryl sarcosinate) and incubating at room temperature for 15 mm. After adding 70 L 5 M NaC1 and 55 uL 8.6 per cent CTAB (cetyltrimethylammonium bromide) and heating at 60°C for 10 mm, the homogenate was extracted with 600 L chloroform:isoamyl alcohol (24:1) and centrifuged in a Kubota KM-15200 microcentrifuge at 5000 g for 10 mm. The nucleic acid was precipitated from the aqueous phase by adding 400 tL isopropanol and pelleted by centrifugation at 12000 g for 10 mm in the microcentrifuge, then washed with 70 per cent absolute ethanol. The pellet was dried and dissolved in 100 1iL TE buffer (10 mrvi Tris—HC1, pH 8.0, 1 mM disodium EDTA), containing 20 mglmL RNase, and stored at — 20°C. The DNA concentration was determined using a Hoeffer fluorometer and adjus- ted to 10 ng/jiL for use in the polymerase chain reaction (PCR). Polymerase chain reaction PCR conditions for RAPD reaction with the Idaho Air Thermal Cycler are described as follows. Each sample, comprising 50 mM Tris—HC1 buffer (pH 8.5) containing 20 mivi KCI, 1.5 mrvt MgCl2, 0.5 mg/mL The Genetical Society of Great Britain, Heredity, 77, 388—395. LOW GENETIC VARIATION IN AMENTOTAXUS FORMOSA NA 391 Table 1 Sequences and codes of random primers and the number of monomorphic and polymorphic fragments amplified Sequence (5' —3') Primer OPE-2 OPE-12 OPE-17 OPE-19 OPS-1 OPS-lO OPS-13 OPS-18 OPY-2 OPY-7 OPY—9 OPY-lO OPY-17 P-4 P-6 P-b P-li P-13 P-14 P-25 No. of monomorphic fragments GGTGCGGGAA TFATCGCCCC CTACTGCCGT ACGGCGTATG CTACTGCGCT ACCGTr'CCAG GTCGTTCCTG CTGGCGAACT CATCGCCGCA AGAGCCGTCA AGCAGCGCAC CAAACGTGGG GACGTGGTGA CGAAGCTTCG CCGTCGACGA ATTGCGTCCA ATGTCCTCGA TCAGCGTGCT TACCGAACGT GGTACCGTGC Total 15 12 10 6 2 0 3 0 7 7 10 1 8 3 13 14 9 21 5 13 1 7 0 2 0 14 15 19 13 1 1 10 9 8 0 0 0 8 19 8 6 5 19 229 (79.2%) Total No. of polymorphic fragments 60 (20.8%) 15 23 5 18 12 17 7 16 15 20 14 13 19 10 9 27 14 289 BSA, 200 tM each of dATP, dCTP, dGTP, dTTP, 0.4 M 10-base primer, 60 ng of template DNA and electrophoresis in 1 x TBE buffer, and detected by means of ethidium bromide staining, viewed under 1.7 units of Taq DNA polymerase (Boeringer ultraviolet light. Specific amplification products were Mannheim Biochemica) at a final volume of 20 1iL, scored as present (1) or absent (0) in each DNA was heat-sealed in a 25 jiL glass capillary tube. Twenty random primers, 13 (OPE-2, ..., OPY-17) supplied by Operon Technologies and 7 (P-4, sample and similarity coefficients (SC) were estima- ted using Nei & Li's (1979) matching coefficient P-25) synthesized by Oligos Etc., were included in the survey (Table 1). The amplification conditions included a total of 45 cycles with template denaturation at 94°C for 60 s, primer annealing at 37°C for 7 s, and primer extension at 72°C for 70 s during the first two cycles. The time for template denaturation was then reduced to 1 s for the remaining 43 cycles. Reactions were further incubated at 72°C for 4 mm SC = 2NABI(NA+NB), method where NA is the number of bands in individual A, NB is the number of bands in individual B, and NAB is the number of bands present in both A and B. Within-population similarity (5) was calculated as the mean of SC across all possible comparisons and the capillaries were stored at 4°C before the between individuals within a population. Betweenpopulation similarity, corrected for within-popula- amplification products were analysed by gel tion similarity, was electrophoresis. S=1+S'—0.5 (S,+S1), Analysis of PCR products products were separated using 1.5 per cent NuSieve 3:1 agarose (FMC BioProducts) gels by PCR The Geneticaf Society of Great Britain, Heredity, 77, 388—395. where S, and S1 are the values of S for population i and j, respectively, and S ' is the average similarity between randomly paired individuals from populations i andj (Lynch, 1990). 392 C.-T. WANG ETAL. Table 2 Allele frequencies and the expected (He) and observed (H0) heterozygosities of the polymorphic locus in the two populations of Amentotaxus form osana Locus and allele Population Tawu Tsatsayalai Avg. a 0.090 b 0.910 0.000 1.000 0.045 0.955 H0 0.180 He 0.164 0.000 0.000 0.090 0.086 Avg. H0 Avg. H0 0.009 0.000 0.000 0.0045 0.004 Table 3 The percentage of polymorphic loci, the percentage of heterozygous loci per individual, the mean number of alleles per locus, and the effective number of alleles per locus for each population of Amentotaxus formosana Tawu 0.008 0 5.0 % Polymorphic Pgi-1 Tsatsayalai loci* %Heterozygous loci/individualt Mean no. of alleles/locust Effective no. of alleles/locus 0.009 (0.009) 0 (0.000) 1.05 (0.05) 1.00 (0.00) 1.01 1.00 Expected heterozygosity for each population was calculated as the arithmetic mean at the 20 loci. *The frequency of the common allele is <0.99. tSE is shown in parentheses. Results Table 4 Results of the y contingency test and F-statistics for Pgi-1 in the two populations of Amentotaxus formosana Isozyme analysis Isozyme patterns from gametophyte and leaf tissue were compared to define the enzyme loci. With the exception of PGI, no differences in band number 7 -lc U.I. 0.804t 1 2 Pgi-1 were found between the gametophyte and leaf Average tissue. The number of loci was determined according to Weeden & Wendel (1989). tNot significant (5%). Ten enzyme systems, with a total of 20 putative loci, were stained with consistently good resolution: two loci for PGM, PGI, IDH, AAT, 6PGD and SOD; three for MDH and EST; and one locus for SKDH and F-EST. All loci were monomorphic, with the exception of Pgi-1, which resolved two cathodally migrating alleles. The observed allele frequencies, observed and expected heterozygosities at the polymorphic locus, and average heterozygosities at the population level are listed in Table 2. Allele Pgi-la was found only in the Tawu population, whereas Pgi-lb was observed in both populations. Two genotypes, ab and bb, have been observed so far. The proportion of polymorphic loci, percentage of heterozygous loci per individual, the mean number of alleles per locus, and the effective number of alleles per locus were 5 per cent, 0.9 per cent, 1.05, and 1.01, respectively, for the Tawu population, whereas no variation was found for Tsatsayalai population (Table 3). F-statistics are listed in Table 4. The X2-test was 11s v r 'IT u 'ST —0.099 —0.099 —0.047 —0.047 0.047 0.047 genotypes within a population was in Hardy—Wein- berg equilibrium. Treating the entire species as a random mating unit, estimates of F11 are closer to zero than F15 for the locus surveyed. The extent of genetic differentiation among populations (FST) was 0.047. Thus, more than 95 per cent of the genetic variation resided within a population. When the 110 individuals originating from the Tawu population were divided into four categories, based on their DBH, it was found that the young cohort with a DBH less than 5 cm had the highest heterozygosity (H = 0.25), whereas the older cohorts, with 15-25 cm and 25 cm DBH, gave lower values of H=0.105 and H=0.111, respectively (Table 5). The negative values of the fixation index indicated that the observed distribution of genotypes within a category had a slight excess of heterozygotes. performed according to the formulae of Li & RAPD analysis Horvitz (1953). The F15 value for the Pgi-1 locus was The 20 random primers used in this study generated negative (—0.099), but the x2 analysis showed no a total of 289 DNA fragments (Table 1). Sixty of these fragments (20.8 per cent) were polymorphic, and 229 (79.2 per cent) monomorphic. The number significant deviation from zero at the 5 per cent level, indicating that the observed distribution of The Genetical Society of Great Britain, Heredity, 77, 388—395. LOW GENETIC VARIATION IN AMENTOTAXUS FORMOSANA 393 Table 5 Genotype distribution and allele frequency at Pgi-1 in four DBH classes in the Tawu population of Amentotaxus formosana 5 5—45 15—25 >25 40 42 19 9 aa 0 10 0 9 0 2 0 ab bb 30 33 17 8 a 0.125 0.875 0.107 0.893 0.053 0.947 0.056 0.944 0.100 0.900 0.250 0.222 0.214 0.194 —0.120 0.105 0.102 0.111 0.200 0.111 —0.059 0.181 DBH(cm) No. Genotype Allele b H0 H, Ft —0.143 —0.056 Mean* 1 *weighted by the number of individuals. tFixation index. M 2 3 4 5 6 7 8 9 M 11 12 13 14 15 16 17 18 19 M 2645 bp 1605 bp 1198 bp Fig. 2 RAPD polymorphism in Amentotaxus formosana using P-14. Lanes 2—9 represent eight individuals from the Tsatsayalai population; lanes 11—19 represent nine individuals from the Tawu population; M represents pGEM DNA size markers. 676 bp 517bp 460 bp 396 bp 250 bp of scorable RAPD fragments generated per primer varied between five and 27, while the number of polymorphic bands per primer ranged between one and 19 (Fig. 2). The size of the DNA fragments ranged between 300—3000 bp. Seven of the primers (i.e. OPE-1; UPS-i; OPY-7; OPY-iO; P-b; P-il; and P-i3) detected no variation and the monomorphic profiles they amplified were shared by all individuals in both populations. Observing the pairwise similarity coefficient (SC) across all possible comparisons, the maximum value of SC (0.992) was found within the Tawu population, and the minimum value (0.939) between the two populations. The average similarity coefficients within the Tawu and Tsatsayalai populations and between them were 0.974, 0.970 and 0.966, respectively (Table 6). The between-population similarity (S,), corrected for within-population similarity, was The Genetical Society of Great Britain, Heredity, 77, 388—395. 0.994, and the corresponding genetic distance between the two populations was 0.006. Discussion The low genetic diversity detected in A. formosana during this study is most likely the result of the geological history of Taiwan. Strong tectonic activities (Penglai orogeny) were recorded in the middle of the Pleistocene (Teng, 1987). Several drastic vegetational changes were recorded in Taiwan during the Pleistocene and the last 60000 years (Tsukada, 1967). The coldest climate prevailed in the Tali glacial age or early Würm glacial age, when a rapid expansion of the boreal elements took place (Tsukada, 1966). It is hypothesized that there was a drastic reduction in the number of trees in Taiwan during this geological age, forming a bottleneck that 394 C.-T. WANG ETAL. Table 6 Average similarity coefficients (SC) within and between populations of Amentotaxus form osana Within population Between populations Tawu Tsatsayalai Not corrected Corrected Similarity* 0.974 (0.010) 0.970 (0.012) 0.966 (0.007) 0.994 *SE is shown in parentheses. resulted in low genetic variation. Species such as A. formosana probably survived the extreme climate fluctuation by migrating to lower-elevation refugia during the Quaternary (Li, 1955). The geological reason for the genetic depauperation of A. formosana may be similar to that which caused the low genetic diversity in red pine; this low diversity resulted from passage through a genetic bottleneck during glacial episodes of the Holocene (Fowler & Morris, 1977; Simon et al., 1986). The low genetic diversity could also be a result of the small populations of A. fomiosana confined to southern Taiwan. Because random genetic drift occurs particularly in small populations (Hartl, 1980), it results in fixation of alleles after many generations. Genetic heterogeneity is often attributed to a local adaptation to environmental variations (Hamrick et al., 1992). The FST value indicates that 4.7 per cent of the genetic diversity found in this study occurred between the two populations. This low interpopulational differentiation is consistent with data from many other conifers (Hamrick et al., 1992). The slight but not significant excess of heterozygotes in the Tawu population (F15 = —0.099) is probably because no individuals with genotype Pgilaa were found. Indeed, Pgi-laa is probably absent from the whole population as a total of 110 individuals, which comprises approx. 16 per cent of the Tawu population, was screened. However, as stated in 'Materials and methods', sampling was not abso- resulted from random genetic drift occurring in a small population. Alternatively, limiting ecological factors, including altitude, moisture, microclimate and their interactions found in natural habitats, suggest a strong selection pressure against Pgi-laa. However, these two hypotheses do not account for the absence of the allele Pgi-la in the Tsatsayalai population, which could be considered as a single panmictic unit with the Tawu population. Amentotaxus formosana was occasionally found between these two populations even if it is uncommon. As outcrossing wind-pollinated gymnosperms have the least variation among populations (Hamrick & Godt, 1990), the exchange of gametes between these two populations is always possible. The occurrence of the allele Pgi-la may also be the result of a recent mutation in the Tawu population. The frequency of allele Pgi-la increased from 0.056 in older trees to 0.125 in the youngest trees. This observation tends to support this hypothesis, as it may explain the absence of allele Pgi-la in the Tsatsayalai population. The increase in frequency of allele Pgi-la may have been caused by selection acting directly upon trees carrying this allele. The allele Pgi-la may eventually be detected in the Tsatsayalai population, as no barrier has been found between them. However, dispersion of this allele may be slow because of the large and heavy seeds, which fall around the mother tree. The flowers may receive predominantly the pollen from nearby relatives, even though the pollen could also be transferred a long distance by wind. The percentage of polymorphism detected using RAPDs was greater than that for isozyme markers. Unfortunately, no RAPD marker specific to either the Tawu or the Tsatsayalai population was found. However, the lack of variation between individuals within and between populations, as revealed by RAPD analysis, agrees with the low level found using isozymes. A similar observation, based on isozyme and RAPD data has been reported for red pine (Mosseler et al., 1992). lutely random owing to inaccessibility. The chance of Acknowledgements allele Pgi-]a not being picked up was always We the Tsatsayalai population is the result of sample (National Taiwan Normal University) for his helpful advice and for stimulating discussion. This research was supported by grant 84AST-2,3-FOD-04 from the Council of Agriculture, Taiwan. possible, given that it is a rare allele. Also the possibility exists that the apparent absence of the allele in size. The increase in frequency of allele Pgi-la in individuals with decreasing DBH (Table 5) may have several explanations. First, the allele Pgi-la may be lost in the Tsatsayalai population but has remained unfixed in the Tawu population; this could have would like to thank Prof. Shong Huang References W. M. AND PITEL, J. A. 1984. Techniques for starch gel electrophoresis of enzymes from forest tree CHELIAK, The Genetical Society of Great Britain, Heredity, 77, 388—395. LOW GENETIC VARIATION IN AMEN TO TA XUS FORMOSA NA 395 species. Can. Forestry Sen.'. Inf Rep. PI-X-42, pp. 19—45. Petawawa National Forestry Institute. 3321—3323. CROW, J. F. AND KIMURA, M. 1970. An Introdution to Popu- NE!, M. 1975. Molecular Population Genetics and Evolution. American Elsevier, New York. lation Genetics Theory. Harper and ROW, New York. FARJON, A., PAGE, C. N. AND SCHELLEVIS, N. 1993. A NE!, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. preliminary world list of threatened conifer taxa. Biodi- Genetics, 89, 583—590. NE!, M. AND U, W. H. 1979. Mathematical model for study- versity and Conservation, 2, 304—326. FERET, i'. P. 1971. Isozyme variation in Picea glauca (Moench) Voss seedlings. Silvae Genet., 20, 46—50. FOWLER, D. P. AND MORRIS, R. W. 1977. Genetic diversity in red pine: evidence for low heterozygosity. Can. I Forest Res., 7, 341—347. HAMRICK, J. L. AND GODT, M. J. W. 1990. Allozyme diversity in plant species. In: Brown, A. H. D., Clegg, M. T., Kahler, A. L. and Weir, B. S. (eds) Plant Population Genetics, Breeding, and Genetic Resources, pp. 43—63. Sinauer Associates, Sunderland, MA. HAMRICK, 1. L., GODT, M. J. W. AND SI-IERMAN-BROYLE, S. L. 1992. Factors influencing levels of genetic diversity in woody plant species. New Forests, 6, 95—124. HARTL, D. L. 1980. Principles of Population Genetics. Sinauer Associates, Sunderland, MA. HEUN, M., MURPH, J. P. AND PHiLIPS, T. D. 1994. A compari- son of RAPD and isozyme analyses for determining the genetic relationships among Avena sterilis L. accession. Theor AppI. Genet., 87, 689—696. LI, C. C. AND HORVITZ, D. 0. 1953. Some methods of esti- ing genetic variation in terms of restriction endonucleases. Proc. Nat!. Acad. Sci. USA., 76, 5269—5273. SIMON, J. P., BERGERON, Y. AND GAGNON, D. 1986. Isozyme uniformity in populations of red pine (Pinus resinosa) in the Abitibi Region, Quebec. Can. J. Forest Res., 37, 10—14. SWOFFORD, D. L. AND SELANDER, R. B. 1989. B!OSYS-1. A Computer Program for the Analysis of Allelic Variation in Population Genetics and Biochemical Systematics. Release 1.7. University of Illinois, Urbana, IL. TENG, L. s. 1987. Stratigraphic records of the Late Cenozoic Penglai orogeny of Taiwan. Acta Geol. Taiwan, 25, 205—224. TSUKADA, M. 1966. Late Pleistocene vegetation and climate in Taiwan (Formosa). Proc. Nati. Acad. Sci. US.A., 55, 543—548. TSUKADA, K!. 1967. Vegetation in subtropical Formosa during the Pleistocene glaciations and the Holocene. Palaeogeogr Palaeoclim. Palaeoecol., 3, 49—64. mating the inbreeding coefficient. Am. J. Hum. Genet., . 1989. Genetics of plant isozymes. In: Soltis, D. E. and Soltis, P. S. (eds) Isozyme 5, 107—117. in Plant Biology, pp. 46—72. Dioscorides Press, Portland, LI, H. L. 1955. The genetic affinities of the Formosan flora. Proc. 8th Pacif Sci. Congr, 4, 189—195. LYNCH, M. C. 1990. The similarity index and DNA fingerprinting. Mol. Biol. Evol., 7, 478—484. MAAI3, H. I. AND KLAAS, M. 1995. Infraspecific differentia- tion of garlic (Allium sativunl L.) by isozyme and RAPD markers. Theor Appi. Genet., 91, 89—97. MEIJER, G., MEGNEGNEAU, B. AND LINDERS, E. G. A. 1994. Variability for isozyme, vegetative compatibility and RAPD markers in natural populations of Phomopsis subordinaria. Mycol. Res., 98, 267—276. MOSSELER, A., EGGER, K. N. AND HUGHES, G. A. 1992. Low levels of genetic diversity in red pine confirmed by random amplified polymorphic DNA markes. Can. J. Forest Res., 22, 1332—1337. MURRAY, M. G. AND THOMPSON, W. F. 1980. Rapid isolation WEEDEN, N. F. AND WENDEL, j. OR. WILLIAMS, J. 0. K., KUBELIK, A. R., LIYAK, K. J., RAFALSKI, J. A. AND TINGEY, s. v. 1990. DNA polymorphisms ampli- fied by arbitrary primers are useful as genetic markers. Nucl. Acids Res., 18, 653 1—6535. WOLFF, K. 1991. Analysis of allozyme variability in three Plantago species and a comparison to morphological variability. Theor App!. Genet., 81, 119—126. WRIGHT, S. 1969. Evolution and the Genetics of Populations, vol. 2, The Theory of Gene Frequencies. University of Chicago Press, Chicago. YANG, S. Z. Studies on the vegetation ecology of Tsatsayalai Shan nature reserve for Amentotaxus formosana in southern Taiwan. Q. J. Chin. For., 27, 19—24 (in Chinese). YEH, C. L., CHEN, C. C., CHUNG, Y. L. AND FANG, 0. T. 1992. of high molecular weight plant DNA. Nucl. Acids Rex., Using GIS techniques to investigate changes of the 8, 4321 —4325. Amentotaxus formosana population. Remote Sensing, NE!, M. 1973. Analysis of gene diversity in subdivided populations. Proc. Nail. Acad. Sd. USA., 70, The Genetical Society of Great Britain, Heredity, 77, 388—395. 28—5 1 (in Chinese). 16,