HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via ISI Web of Science (1) Citing Articles via Google Scholar Google Scholar Articles by Gottlieb, L. D. Articles by Ford, V. S. Search for Related Content PubMed PubMed Citation Articles by Gottlieb, L. D. Articles by Ford, V. S. Agricola Articles by Gottlieb, L. D. Articles by Ford, V. S.

The status of Clarkia australis (Onagraceae)¹

Section of Evolution and Ecology, Division of Biological Sciences, University of California, Davis, California 95616

Received for publication February 27, 1998. Accepted for publication August 13, 1998.

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Clarkia australis and C. virgata grow on the western slope of the central Sierra Nevada of California. Clarkia australis was established to accommodate populations of C. virgata from south of the Tuolumne River that could not be successfully hybridized to populations north of the river. Although the species is maintained in the new Jepson Manual, its validity has been questioned because only two populations were originally tested, and they had no useful morphological traits that distinguished them from C. virgata. We report here the results of a large program of interpopulation hybridizations that show that C. australis is distinct and that its reproductive isolation from C. virgata is complete and absolute and reflects a compatibility block that apparently causes abortion of hybrid seeds in early development. Both species include populations north and south of the Tuolumne River and, in general, those of C. australis occupy higher elevations. Morphologically, the species are extremely similar though the mean values of several dimensions of the petals are different. However, significant variation among their populations has the consequence that, at present, the only certain way to assign particular populations to species is to test their compatibility with previously tested populations.

Key Words: Clarkia australis • Clarkia virgata • hybrid incompatibility • Onagraceae • reproductive isolation • speciation

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Prior to the present use of molecular evidence to infer phylogenetic relationships among plant species, the most convincing evidence generally resulted from cytogenetic analyses of interspecific F1 hybrids that revealed chromosomal homologies. In cases in which there was an aneuploid change in chromosome number, it was often possible to ascertain the direction of evolution and show that a Species A gave rise to Species B (Tobgy, 1943 ; Sherman, 1946 ; Lewis and Roberts, 1956 ; Small, 1971a ; Jackson, 1962 ; Kyhos, 1965 ; Smith, 1974 ; Carr, 1975 , 1980 ). Small (1971a) studied the diploid species of Clarkia sect. Myxocarpa, native to the northern and central Sierra Nevada of California. At the start of his analyses, only two diploid species were known in the section: C. mildrediae (n = 7) and C. virgata (n = 5), the progenitors of the allotetraploid C. rhomboidea (n = 12) (Lewis and Lewis, 1955 ; Mosquin, 1964 ). Small determined that northern populations of C. mildrediae were reproductively isolated and he named them C. borealis. He found that some populations of C. virgata had n = 6, leading to their designation as C. mosquinii, and also that other, mostly southern, populations of C. virgata were reproductively isolated so he named them C. australis (n = 5). All of these diploid species are outcrossing and nearly indistinguishable morphologically, but differ by very strong barriers to hybridization and at least two reciprocal translocations and often more. The chromosomal differences resulted in extreme reductions in pollen fertilities of their interspecific F1 hybrids.

The geographical distributions of these species correlate closely with their chromosome numbers: C. borealis and C. mildrediae, in the north, both have n = 7; C. mosquinii, with a central distribution, has n = 6; and C. virgata and C. australis, in the south, both have n = 5. Each of these species has a very limited distribution, generally within one or two counties. Small (1971a ; Fig. 1) suggested that C. borealis gave rise to C. mildrediae and, independently, produced C. mosquinii with which it shares a similar stem/branch architecture. Then, C. mosquinii subsequently gave rise to both C. virgata and C. australis. The latter two species were considered to have originated independently, rather than one from the other, because hybrids between each of them and C. mosquinii could be synthesized, but no hybrids could be made between them. Evolution of extant species one from another is particularly interesting because it provides an opportunity to identify specific chromosome segments and genes that moved from Species A (C. borealis) to Species B (C. mosquinii) but not to Species C (C. mildrediae), and others that may have moved twice, from C. borealis to C. mosquinii and from C. mosquinii to C. virgata and to C. australis.

View larger version (12K): [in this window] [in a new window]   Fig. 1. Phylogenetic relationships of the diploid species in Clarkia sect. Myxocarpa proposed by Small (1971a) .

With the support of the Stanislaus National Forest, which manages the land where these plants are found, we decided to look into the status of C. australis. Its status was important to the Stanislaus National Forest because it is rare and grows in openings and other sites where it may be affected by various logging and management activities such as reforestation following burns or salvage of insect- and drought-killed trees, and thus it may require federal protection. Studies to determine the impact of such activities on the persistence of the species cannot be justified until it is certain that C. australis is valid.

Small tested only two populations from south of the Tuolumne River and neither of them made hybrids with seven populations from north of the river. The failure to make hybrids between closely related Clarkia species is unusual and suggested that his results could reflect poor growing conditions or some peculiarities of the individual plants he tested. Here we report the results of an extensive program of experimental hybridization between numerous pairs of populations collected on both sides of the river. The populations were studied in a uniform greenhouse environment so that their morphological characteristics could be examined under the same growth conditions. A limited electrophoretic analysis of isozyme variability was also done.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Populations examined
Seeds of 40 individuals from each of 18 populations, eight north and ten south of the Tuolumne River, were collected by employees of the Stanislaus National Forest, Groveland Ranger District (Table 1, Fig. 2). The sampled populations included collections made at localities numbered and listed in Small (1971a) with precise geographical description: Small 7 (ICE HOUSE; also LDG 9308); Small 9 (COSUMNES; also LDG 9309); Small 10 (OMO); and Small 14 (MATHER WEST). Small considered his collections 7, 9, and 10 to be C. virgata and his collection 14 to be C. australis.

View larger version (22K): [in this window] [in a new window]   Fig. 2. Map of east-central California showing locations of collections listed in Table 1 . C. australis: 1 = BEA, 2 = CAR, 3 = COT, 4 = FEM, 5 = MAT, 6 = PLU, 7 = TRU, 8 = TUO, 9 = WIL. C. virgata: A = CHI, B = COS, C = CRO, D = GRO, E = HAR, F = HEL, G = ICE, H = OMO, I = PIL. The approximate location of the Tuolumne River is shown in Tuolumne County.

Experimental crossing scheme
The hybridization program was carried out in three successive springs: 1995, 1996, and 1997. Plants from six populations were grown in 1995 and, in 1996, 12 populations, including three previously studied in 1995, were grown. Three additional populations and previously tested collections were grown in 1997. Each year, between three and six plants from each population were crossed reciprocally (as both male and female) to each of the other populations. For each combination tested, at least six flowers on several plants were cross-pollinated. The number of mature enlarged fruits, indicating successful pollination, and the number of seeds in each fruit were determined.

Morphological studies
Visual examination of the morphological appearance of plants of each population showed that all of them were highly similar as Small had reported. Particular attention was paid to leaf shape since he regarded it as the diagnostic of the species. Plants grown from seeds collected from some localities previously studied by Small did exhibit relatively narrow lanceolate-shaped leaves, for example, Small 14, which he had designated C. australis, whereas Small 10, designated C. virgata, had relatively wide elliptically shaped leaves. However, these and many other populations had individuals with both leaf shapes. Leaf shape was also found to vary during the growth of single plants so that individuals scored as having leaves of one shape were frequently scored several weeks later as having leaves of a different shape. Such variability in shape both within and among populations and on single plants makes the character inappropriate and of little taxonomic value in distinguishing the two species.

However, there appeared to be more useful variation in some dimensions of the flower petals, and they were measured in 1996 and 1997 in the Davis greenhouse. Twelve plants were used from each of eight populations of C. australis, eight of C. virgata, and one of C. mosquinii. One petal from each of the first two flowers to open was selected and placed on a 3 x 5 index card under Scotch tape to keep it flat. Each petal was harvested on the day the stigma became receptive, 3–5 d after anthesis. The time of stigma opening is a landmark because petals continue to enlarge until then, but not thereafter. Limb width, "isthmus" width at the narrowest point (the isthmus separates the limb, or distal region, from the claw, or proximal region), claw width, limb length and claw length (Fig. 3) were measured with a micrometer (six lines = 1 mm) in the eyepiece of a binocular microscope. Total length was computed as the sum of limb and claw length. Single classification and nested analysis of variance were used to test for significant differences among taxa, and among and within populations. The data were also analyzed with the BMDP7M stepwise discriminant analysis program (BMDP Statistical Software, 1993) to determine whether some linear combination of measured variables might provide a clearer separation of C. australis from C. virgata.

View larger version (25K): [in this window] [in a new window]   Fig. 3. Diagram of typical petal of Clarkia australis/virgata/mosquinii showing the dimensions measured.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Experimental hybridizations
Table 2 summarizes the results of the crossing scheme. All plants used as females were self-pollinated to confirm fertility, and all self-pollinations were successful. Cross-pollinations in all combinations between WILSON, PLUM FLAT, FEMMONS SOUTH, TUOLUMNE, and COTTONWOOD from north of the river and MATHER WEST, BEAR MTN, CARLON, and TRUMBULL from south of the river yielded enlarged fruits containing a nearly full set of large dark seeds (40–50), indicating successful hybridization. Each fruit also contained some small flakes, presumably unfertilized ovules.

Cross-pollinations between populations representing these two groups also produced enlarged fruits, generally a sign of successful fertilization in Clarkia species. However, these fruits contained few or no full-sized dark seeds but numerous pale, shrunken, intermediate-sized seeds of a type not seen in the crosses within groups, in addition to the small unfertilized flakes. The enlargement of the fruits and immature appearance of the seeds suggested that fertilization was followed by abortion of the developing embryos. For the pollinations done in 1995, parent plants were chosen so that hybrid seeds would be recognizable by electrophoretically separable allozymes. All large, dark seeds produced from these crosses were moistened a few days in petri dishes and examined under a dissecting microscope. Most had no visible embryos and a few had embryos aborted at the heart or torpedo stages. A number of seeds germinated successfully but electrophoretic examination of the seedlings demonstrated that nearly all had resulted from self-pollination. Only 12 seedlings appeared to be true hybrids, one from the cross between CARLON and HARDEN and 11 from the cross between CARLON and PILOT. Ten of these seedlings matured to flower. The plants were small, and their entire epidermis appeared rough and abnormal. The flowers on the hybrid plants had abnormal morphology, including irregular numbers and asymmetries of petals and various adnations between floral organs. The pollen was shrunken and poorly formed. Since one cross-compatible group included MATHER WEST collected from the same site as Small 14, this group is identified as C. australis. The second group of compatible populations included OMO, the same as Small 10, and is identified as C. virgata.

Thus, the hybridization program demonstrated clearly that C. australis and C. virgata are fully reproductively isolated by extreme cross-incompatibility. When this barrier is occasionally breached, the F1 hybrids are morphologically abnormal and completely sterile. The geographical distributions of the two species do not correlate with location north or south of the Tuolumne River. Both species are found both north and south of the river. In general, C. australis occupies higher elevation sites from 1250 to 1524 m, with one population at 1036 m, and C. virgata lower sites between 549 and 1265 m, with one population at 1615 m (Table 1). Both species are found on the same soil substrate, primarily metasedimentary granitic sandy loam.

Analysis of petal dimensions
Descriptive statistics for petal dimensions are shown (Table 3) for both species, as well as for a single population of C. mosquinii. Single classification analysis of variance indicates significant differences (P < 0.001) in the mean values of C. australis and C. virgata for all the traits measured. However, most measurements for each species fall in the range of the other. Petal dimensions of their putative parent C. mosquinii also largely fall in the same range.

View larger version (14K): [in this window] [in a new window]   Fig. 4. Mean coordinates of populations of C. australis and C. virgata (symbols as in Fig. 2 ) and C. mosquinii (solid square) plotted on the first two canonical axes, along with individual measured petals of C. australis (solid dots) and C. virgata (open dots) or both (+).

Since no mixed populations have been found, it was worth considering the utility of representing a population by the mean values of the measured characters rather than by individual measurements. A discriminant analysis based on population means, using the two species as groups, produced a single canonical variable that successfully separated the populations of C. australis (with positive values) from those of C. virgata (negative values; histogram not shown). However, the canonical value for the PLUM FLAT population of C. australis, 0.05, was barely different from zero, whereas other populations of C. australis had values from 1.74 to 3.93.

The population of C. mosquinii that was examined was roughly between its two descendant species, though somewhat closer to C. australis, consistent with Small's suggestion that the two species with n = 5 originated independently.

Petals were measured for the first two flowers to open on each plant. A paired t test showed the two measurements were not significantly different except that limb length was slightly longer (11.04 vs. 10.91 mm, F = 6.9, df = 1,191, P < 0.01) on the second flower.

Electrophoretic comparison
A small study was initiated to select electrophoretically distinguishable plants to use as parents in the test crosses done in 1995 and to determine whether individuals in the field might be identified to species by genes encoding allozyme variants. The isozymes examined appear to be encoded by nine genes. The four surveyed populations of both species had identical electrophoretic mobilities of PGIP (plastid enzyme), MDH1, MDH2, GOT, IDH1, and IDH2. The latter three isozymes were only examined in the CHINA FLAT and WILSON populations. Three isozymes were polymorphic: PGIC (cytosolic), esterase, and MDH3 (Table 5). In general, for each locus, the same one or two alleles have high frequencies in both species. At MDH3, no variability was found in C. australis, but this may be because only two populations were surveyed. At this locus, the high-frequency allele of C. virgata is the same as the allele fixed in C. australis. The electrophoretic results show some differences in allele presence/absence and frequency, but it remains to be determined whether this type of evidence will be useful to distinguish the species.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The present study demonstrates convincingly that C. australis and C. virgata are distinct species. The evidence comes from a large program of interpopulation hybridizations showing that there are two groups of populations, within each of which all populations are fully interfertile but between which reproductive isolation is complete and absolute. Both species include populations north and south of the Tuolumne River. Clarkia australis occupies sites that are generally but not always higher in elevation and, in fact, the highest elevation population tested, CROCKER at 1615 m (Table 1) proved to be C. virgata. The lowest population of C. australis, PLUM FLAT at 1036 m, was lower than half the tested populations of C. virgata.

The remarkable similarity of the two species presently prevents assignment of populations to one or the other on the basis of morphology. Although the means of several petal dimensions differ significantly between the species and discriminant functions were 95% successful in distinguishing the species in our study, the significant variation among populations makes it unlikely that discriminant functions from the present analysis could be applied to identify newly discovered populations. For example, if the PLUM FLAT population of C. australis is not used in the discriminant analysis, the resulting discriminant functions misclassify 21 out of 24 PLUM FLAT petals as C. virgata and would certainly result in the misidentification of the population. Similarly, if PLUM FLAT is omitted from a discriminant analysis based on population means, the resulting classification function identifies it as C. virgata. PLUM FLAT is the only such discordant population presently known (Fig. 4), but it would not be appropriate to assume that others will not be found.

A probable additional disadvantage to the use of petal dimensions to identify populations in the field is the greater variability of field measurements compared to the study measurements, which were done with an ocular micrometer on petals harvested at maximum size from plants grown in an equable and uniform greenhouse environment. On these small plants, petal size declines after the first few flowers, further increasing the variability to be expected from a random sample.

The limited electrophoretic analysis suggests that although populations are likely to differ in allele presence/absence and frequency there may not be any definitive electrophoretic test that will correctly assign populations to species. Thus, at the present time, the only certain way to assign particular populations to species is to test their compatibility with previously tested populations. The geographical distributions shown on Fig. 1 are certainly incomplete and may even be misleading in suggesting that C. australis has a smaller range than C. virgata. The difficulty of circumscribing the species' geographical distributions is further compounded by the fact that both have often been confused in the field with the tetraploid C. rhomboidea, which occurs throughout the area.

The lack of a qualitative morphological difference between two species is unusual in Clarkia. The vegetatively highly similar sister species C. biloba and C. lingulata are readily identified by their bilobed or strap-shaped petals (Lewis and Roberts, 1956 ). The similar C. jolonensis and C. bottae can be distinguished by differences in the color and surface (scales vs. papillae) of their seeds (Parnell, 1970 ). Even C. borealis and C. mildrediae, lacking qualitative differences, are evidently different in stem architecture (Small, 1971b ). However, the two subspecies of C. tembloriensis differ morphologically only in the lengths of certain floral characters (Holsinger, 1985 ).

Complete failure of hybridization between closely related species of Clarkia is also unusual because in general the species can be hybridized, though the hybrids are nearly always sterile. The present results are reminiscent of those reported by Parnell (1968) . He showed that the morphologically very similar C. jolonensis, from Monterey County, and C. bottae (formerly C. deflexa) in the Coast Range to the south were reproductively isolated because crosses between them also produced hybrid embryos that failed to develop beyond early stages. Failure of hybrid embryos to develop has also been found in Oryza (Chu and Oka, 1970 ), Melilotus (Sano and Kita, 1978 ), and Mimulus (Vickery, 1959 , 1978 ).

It has not been possible during the present study to compare the ecological properties of the two species. Additional study from this point of view may reveal important differences in associated species, or in germination requirements, or in climatic factors such as differences in rainfall or temperature regimes. The difference in average elevation of their populations suggests such analysis would be worthwhile.

In conclusion, our results support and extend Small's (1971a) finding that C. australis is distinct from C. virgata and is more widespread than previously realized. We show that reproductive isolation between them follows from a compatibility block that apparently causes hybrid seeds to abort in early stages of development. The complete sterility of the very few hybrid individuals that flowered most likely reflects substantial differences in chromosomal arrangement as Small (1971a) found in hybrids between other species in the section. Yet to be determined is whether the two species originated independently as he suggested from their n = 6 progenitor or one from another. The overall similarity of C. australis and C. virgata to each other and to C. mosquinii is consistent with either possibility and suggests recent origins. Their derivation will probably best be examined by phylogenetic analysis of nucleotide sequences, and such a study can be expected also to provide information about gene sorting after population bottlenecks since a recent aneuploid change in their ancestry is certain.

	FOOTNOTES

 
¹ This study was supported by a Challenge Cost-Share Agreement between the Stanislaus National Forest and the University of California, Davis.

² Author for correspondence.

	LITERATURE CITED

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Carr, G. D. 1975 Chromosome evolution and aneuploid reduction in Calycadenia pauciflora (Asteraceae). Evolution 29: 681–699.[CrossRef][ISI]

———. 1980 Experimental evidence for saltational chromosome evolution in Calycadenia pauciflora Gray (Asteraceae). Heredity 45: 107–112.[ISI]

Chu, Y-E., and H-I. Oka. 1970 The genetic basis of crossing barriers between Oryza perennis subsp. barthii and its related taxa. Evolution 24: 135–144.

Holsinger, K. E. 1985 A phenetic study of Clarkia unguiculata Lindley (Onagraceae) and its relatives. Systematic Botany 10: 155–165.

Jackson, R. C. 1962 Interspecific hybridization in Haplopappus and its bearing on chromosome evolution in the Blepharodon section. American Journal of Botany 49: 119–132.[CrossRef][ISI]

Kyhos, D. W. 1965 The independent aneuploid origin of two species of Chaenactis (Compositae) from a common ancestor. Evolution 19: 26–43.[CrossRef][ISI]

Lewis, H. 1993 Clarkia. In J. C. Hickman [ed.], The Jepson Manual, 786–793. University of California Press, Berkeley, CA.

———, and M. E. Lewis. 1955 The genus Clarkia. University of California Publications in Botany 20: 241–392.

———, and M. R. Roberts. 1956 The origin of Clarkia lingulata. Evolution 10: 126–138.

Mosquin, T. 1964 Chromosomal repatterning in Clarkia rhomboidea as evidence for post-pleistocene changes in distribution. Evolution 18: 12–25.

Parnell, D. R. 1968 Reproductive barriers in Clarkia deflexa. Brittonia 20: 387–394.

———. 1970 Clarkia jolonensis (Onagraceae), a new species for the inner coast range of California. Madroño 20: 321–323.

Sano, Y., and F. Kita. 1978 Genes for reproductive isolation located on rearranged chromosomes. Heredity 41: 377–383.[ISI]

Sherman, M. 1946 Karyotypic evolution: a cytogenetic study of seven species and six interspecific hybrids of Crepis. University of California Publications in Botany 18: 369–408.

Small, E. 1971a The evolution of reproductive isolation in Clarkia section Myxocarpa. Evolution 25: 330–346.

———. 1971b The systematics of Clarkia, section Myxocarpa. Canadian Journal of Botany 49: 1211–1217.

Smith, E. B. 1974 Coreopsis nuecensis (Compositae) and a related new species from southern Texas. Brittonia 26: 161–171.[CrossRef][ISI]

Tobgy, H. A. 1943 A cytological study of Crepis fuliginosa, C. neglecta and their Fl hybrid, and its bearing on the mechanism of phylogenetic reduction in chromosome number. Journal of Genetics 45: 67–111.[ISI]

Vickery, R. K. 1959 Barriers to gene exchange within Mimulus guttatus (Scrophulariaceae). Evolution 13: 300–310.[CrossRef][ISI]

———. 1978 Case studies in the evolution of species complexes in Mimulus. Evolutionary Biology 11: 405–506.

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via ISI Web of Science (1) Citing Articles via Google Scholar Google Scholar Articles by Gottlieb, L. D. Articles by Ford, V. S. Search for Related Content PubMed PubMed Citation Articles by Gottlieb, L. D. Articles by Ford, V. S. Agricola Articles by Gottlieb, L. D. Articles by Ford, V. S.