INTERGENERIC HYBRIDS IN THE MEXICAN CRASSULACEAE: II. DUDLEYA (AND PLATE TECTONICS)

CHARLES H. UHL

Plant Biology,

Cornell University,

Ithaca, NY 14853-5908

Abstract.

Probably there are no authentic intergeneric hybrids of Dudleya. Other genera of Mexican Crassulaceae hybridize in cultivation, but earlier reports of intergeneric hybrids of Dudleya are questionable, and all 39 recent attempts were unsuccessful. Most species of Dudleya occur on the Pacific geological plate, west of the San Andreas Fault and the Gulf of California, whereas related genera clearly are all centered on the North American Plate to the east. A speculative scenario proposes that the unique progenitors of Dudleya adapted to coastal habitats on a scramble of mixed bits of land ("terranes") that had earlier been pushed against the leading edge of the westward-moving North American Plate, diverging from the progenitors of its closest relative, Echeveria, as the latter adapted to higher elevations inland. About 27 million years ago, plate movement began to transfer the habitats of Dudleya onto the northward-moving Pacific Plate, but it is not known whether the genus or its progenitors were there yet. Five million years ago Dudleya was certainly aboard when separation of Baja California from the mainland completed its geographic isolation. During the glacial times that began about two million years ago the ocean greatly moderated the climate in the coastal habitats of Dudleya, and it escaped the severe environmental stress and the pressure for evolutionary change that its old relatives on the mainland endured.

Hybrids among the Mexican Crassulaceae are very uncommon in the wild, probably because those few species that occur together and flower at the same time are served by different polli­nators. Nevertheless, most species and genera hybridize easily in cultivation, and I have pro­duced hybrids that connect, directly or indirect­ly, more than 200 species and as many as nine genera into a biosystematic unit called a comparium, one of the largest known (Uhl, 1992b).

Dudleya clearly does not have the same ability to hybridize readily with other genera. Although Rowley (Jacobsen and Rowley, 1958) proposed x Dudleveria as a special name for its hybrids with Echeveria, I am very skeptical that authen­tic hybrids have ever been produced between Dudleya and any of the other genera. Expanding on an earlier report (Uhl, 1992a). this paper sum­marizes why I believe this, with comments re­garding distributions and geological plates, and a speculative scenario for the evolution of Dud­leya.

Some species of Dudleya and Echeveria are similar in appearance, and early botanists put them together in the same genus, at first in Cot­yledon and later in Echeveria. Even after Britton and Rose (1903) separated the genus Dudleya, such authors as Berger (1930) and Kearney and Peebles (1954) still kept its species in Echeveria. However, Dudleya and Echeveria are distinct geographically, cytologically and morphologi­cally.

Chromosomes

All species of Dudleya have a basic chromo­some number of 17 (Uhl and Moran, 1953), and about 35% of the populations are polyploid, up to 16-ploid (n = 136). Within Dudleya, inter­specific hybrids, including some between its subgenera Dudleya, Stylophyllum. and Hasseanthus, occur in nature (e.g., Moran, 1951). although they are uncommon, and the species hybridize with each other readily in cultivation. Diploid hybrids show very good chromosome pairing with no detectable abnormalities at meiosis, and some of these hybrids appear to be fertile. Thus, not only do all dudleyas have the same basic chro­mosome number, but even after the meiotic shuf­fling and reduction processes at least some of their hybrids produce functional reproductive cells. This is possible only if the genes in the corresponding parental chromosomes are ar­ranged essentially in the same order. Apparently any changes in Dudleya's chromosomes during evolution have consisted primarily or entirely of gene mutations at the molecular level, with few or no rearrangements in gene sequences.

The cytological situation of Echeveria con­trasts sharply with that in Dudleya. Echeveria has every gametic chromosome number from 12 to 34, as well as many higher numbers (Uhl, 1992b). The behavior of the chromosomes at meiosis in hybrids shows that many changes in the sequen­tial organization of the genome have occurred during the evolution of Echeveria. The chro­mosomes must have broken many times and the broken ends joined together in many different ways. When these breaks and reunions trans­ferred all the vital genetic material from one chromosome to others, the non-essential rem­nant could be lost, lowering the gametic number by one. Clearly in Dudleya and Echeveria the mechanisms of chromosome change during evo­lution have been very different, and this probably indicates that they are not as closely related as similarities in their general appearance might suggest.

Fig. 1. Dudleya pulverulenta, near its type locality. "St. Diego" (R. Moran).

Supposed Hybrids

During the 1870's J. B. A. Deleuil was a nurs­eryman in Marseilles who produced and sold many hybrids of Echeveria. He gave the name Echeveria spiralis (Deleuil, 1875) to hybrids he had grown from seed following pollination of "E. dectpiens" by "E. californica." Recognizing that "E. californica" is a synonym of Dudleya coty­ledon (=D. caespitosa), Rowley (in Jacobscn and Rowley, 1958) concluded that Deleuil's E. spira­lis was therefore a hybrid between Dudleya and Echeveria, and he proposed the name x Dudleve-ria spiralis for this presumed intergeneric hybrid. Unfortunately this conclusion does not survive closer inspection.

Echeveria decipiens, one of the alleged parents of x Dudleveria spiralis, is a Peruvian plant that Baker (1870) originally named Cotyledon deci­piens. It has short (half-inch) leaves densely ag­gregated all along the stem, not in rosettes as usually in Echeveria; the inflorescence is termi­nal, not lateral as in Echeveria; and the petals are about 8 mm long, united for about 4 mm. then strongly recurved outward. In fact Baker said that "Till it flowers it has every appearance of a Sedum." Even when Morren (1874) trans­ferred it to Echeveria. he said that it was not a proper Echeveria but was more like a Sedum. Finally Jacobsen (in Jacobsen and Rowley, 1958) transferred E. decipiens to Villadia, where it seems more at home. Remarkably, Rowley, in naming x Dudleveria in the same paper (Jacobsen and Rowley, 1958), still called it Echeveria decipiens! Thus, if the parents of Deleuil's hybrid E. spiralis were correctly identified and named, an Eche­veria was not one of them. Since Rowley's x Dudleveria did not have the intergeneric parentage he claimed, presumably the name is an empty category (and therefore not valid?). I am equally skeptical that Dudleya will hybridize with Villadia.

Fig. 2. Drawing said to be of "Echeveria pulverulenta" (=Dudleya pulverulenta). from Morren (1874).

Other putative dudleverias have been men­tioned in the literature. For example, Morren (1877) listed 17 additional named hybrids of Echeveria that Deleuil had shown at an exposi­tion at Liège, Belgium. Six of these hybrids had "E. pulverulenta" (=Dudleya pulverulenta, Fig. 1) as one of their parents, and these hybrids must be considered as possible dudleverias. The ques­tion then becomes whether Deleuil's "E. pulverulenta" was correctly identified. Baker's (1869) figure of Cotyledon pulverulenta flowering at Kew appears to be authentic D. pulverulenta, but he said "the plant is of rare occurrence in collec­tions, owing to the difficulty... of keeping it... healthy" and that his plant had flowered only once during many years in cultivation. Van Houtte (1873) and Morren (1874) (Fig. 2) pub­lished drawings of plants grown as E. pulveru­lenta on the European continent, but neither fig­ure shows the inflorescence and flowers critical for identification and the rosettes do not look quite right: the leaves appear relatively too thin, recurved more toward their tips and possibly with longitudinal grooves. Although rather sim­ilar to D. pulverulenta, the plants shown could well be an echeveria, most likely one of the glau­cous-leaved species of Walther's (1972) series Gibbiflorae. Morren (1874) even commented that van Houtte's (1873) drawing looked different from a figure of the species that Lemaire had published in 1845 but which I have not been able to track down. I doubt very much that true D. pulverulenta was a parent of these hybrids and that they really are hybrids between Dudleya and Echeveria. Until a much better case can be made. I think there is no need for the name x Dudle­veria, even if it were valid.

The conclusion is reinforced by my own lack of success in attempts to hybridize Dudleya with Echeveria and other genera. Close to 50% of all my cross-pollinations, both interspecific and in­tergeneric, have resulted in genuine hybrids, al­though the rate is lower for hybrids of Villadia and Lenophyllum with the other genera. By con­trast, my batting average for hybrids between Dudleya and the other genera is just plain zero. Thirty-nine attempts to cross one or another of eight species of Dudleya with various of the other genera have yielded no hybrids at all. These in­clude 18 attempted crosses with 11 species of Echeveria, 10 attempts with four species of Graptopetalum, four with two species of Cremnophila, four with one species of Thompsonella, and one each with a sedum of subgenus Gormania, a pachyphytum, and a villadia. All but two of the other species had been successfully hybridized with others. Diploid and polyploid species of Dudleya (subgenera Dudleya and Hasseanthus) were cross-pollinated with diploid and polyploid species of the other genera, all without result. As usual, it is virtually impossible to prove a neg­ative, and additional attempts might eventually produce hybrids. However, the high rate of suc­cess in making hybrids among the other genera contrasts sharply with the lack of success in mak­ing hybrids of Dudleya. and this demonstrates convincingly that Dudleya is further removed reproductively from the other genera than they are from each other.

Fig. 3. Map of part of western North America. Dotted lines and arrows indicate boundaries of tectonic plates and directions of their movement. Some localities and distributions of species indicated by letters: A = Sedum alamosanum; D = Dudleya saxosa ssp. collomiae and D. pulverulenta ssp. arizonica; F = D. farinosa; G = Gulf of California; N = Sedum niveum; P = East Pacific Rise; S = San Andreas Fault; V = Puerto Vallarta.

Geographic Isolation

Dudleya has apparently been geographically isolated from the other genera on a different geo­logical plate for a long time, and so this repro­ductive isolation is perhaps not surprising. The San Andreas Fault (S in Fig. 3) very closely par­allels the coast of California from Cape Mendocino south to near San Francisco, where it turns more inland toward the southeast, passing north and east of the Los Angeles Basin and just east of San Bernardino and the Salton Sea. from where parallel faults continue down the Gulf of California (G in Fig. 3). In terms of plate tec­tonics the Fault, Gulf line marks the present boundary between the Pacific Plate to the west and the North American Plate to the east (Redfern, 1983). Dudleya clearly is centered and based on the Pacific Plate, west of this Fault/Gulf boundary, whereas the other genera are all based on the North American Plate, east of this bound­ary. Exceptions to this generalization are re­markably few, and I think that they do not de­tract from its validity.

Dudleya farinosa (F in Fig. 3) ranges from its type locality near Monterey, on the Pacific Plate, north along the coast to extreme southern Oregon, on the North American Plate. Forms and relatives of D. cymosa extend into the Coast Ranges and Sierra Nevada, and subspecies of D. saxosa and D. pulverulenta (D in Fig. 3) occur as far east as central Arizona. where they overlap the ranges of Graptopetalum bartramii and G. rusbyi. Probably all of these dudleyas are recent immigrants to these areas. Clearly the great con­centration of species and variation of Dudleya is on the Pacific Plate west of the Fault/Gulf bound­ary, where perhaps more than 40 species occur.

From the opposite direction, several species of Sedum have a few populations on the Pacific Plate west of the Fault Gulf line, but most of their populations and all of their relatives are to the east. Moran (1969) reported Sedum alamosanum (A in Fig. 3) at two localities in the moun­tains of Baja California, but it clearly is centered in Sonora, on the mainland (Uhl, 1985). Sedum niveum (N in Fig. 3) has diploid forms (n = 16) on the North American Plate high in the New York Mountains near the Nevada state line and very near the Fault in the San Bernardino and Santa Rosa Mountains of southern California, and it has octoploids (n = 64) in the Sierra San Pedro Martir in northern Baja California on the Pacific Plate, but its closest relatives (e.g.. S. cockerellii) are all east of the Fault/Gulf line (Moran, 1969; Uhl, 1985). The type locality of Sedum radiatum is just west of the plate bound­ary near Monterey, and one or two other pop­ulations occur west of the Fault, but it is widely distributed east of the Fault in central and north­ern California, and all of its relatives are on the North American Plate (Uhl, 1977). Sedum spathulifolium, ranging from British Columbia to southern California, has several populations west of the Fault, but again these appear to be recent immigrants. The Crassulaceous flora of North America clearly differs on the two plates. The differences between the floras of the plates cor­respond to the biological differences shown by the ability to hybridize or not.

Plate Tectonics

The geological history of the region is certainly related to these differences, but many details re­main uncertain or unknown. Three geological plates are involved (Redfern, 1983). Lines of seafloor spreading, the Eastern Pacific Rise (P in Fig. 3), mark the boundary between the Pacific Plate, which is moving to the north and west, and the Farallon Plate, which is moving away from it toward the east and south. The North American Plate is moving to the west and has overridden most of the old Farallon Plate, leav­ing only the Juan de Fuca Plate (off the coast from southern British Columbia south to north­ern California) and the Rivera and Cocos plates (off the coast of central and southern Mexico and southward) as its only remnants that remain un­covered today (Fig. 3).

As the North American Plate moved to the west it scraped bits of land, called suspect terranes, from the surface of the subducting Far­allon Plate and added them to its leading edge (Redfern, 1983; Sullivan, 1984). Suspect terranes consist of a jumble of rocks of many kinds having different ages and histories, and they originated at different times from many different, scattered sources at various distances from where they now lie. Some of them apparently were old islands or bits of sea bottom that earlier had been carried eastward on the Farallon Plate until at various times they were crushed against the overriding North American Plate and the direction of their movement was reversed (Atwater, 1989).

About 27 million years ago, during the Oligocene epoch, westward movement of the North American Plate began to push the suspect ter­ranes along its leading edge across the East Pacific Rise, where eastward movement of the Farallon Plate had originated, and onto the Pacific Plate (Atwater, 1989). Contact began in the area near the present United States-Mexican border and spread first to the north. The Pacific Plate is mov­ing in a different direction, and relative move­ment along its contact with the North American Plate is more lateral and sliding. The San An­dreas Fault forms this boundary in California, and spasmodic slipping along this and associated faults produces most of the severe earthquakes there. The Pacific Plate is generally estimated now to be moving north at about 5 cm (2 inches) a year, or a mile in 30.000 years, but earlier rates may have been different.

About 16 million years ago contact between the North American and Pacific plates began to extend rapidly southward from its initial point, reaching what is now the tip of the peninsula by about 12.5 million years ago (Atwater, 1989). The East Pacific Rise became inactive as the Far­allon Plate in that region was covered, and Baja California began to be pushed onto the Pacific Plate. Then, about 5.5 million years ago (late Miocene), short new sectors of sea-floor spread­ing developed farther east, alternating in zig-zag fashion with long, sliding, strike-slip sectors (At­water, 1989). This reactivation of the East Pacific Rise tore Baja California away from the main­land and opened the Gulf of California (G in Fig. 3).

Geologists generally believe that those suspect terranes first thrust across the Fault onto the Pa­cific Plate have been carried about 500 km (300 miles) to the north and west since this movement began about 27 million years ago (Atwater, 1989). The shape of the present coast suggests that the tip of the Baja California peninsula may origi­nally have lain about 500 km to the southeast, near Puerto Vallarta (V in Fig. 3). However, only about 300 km of this movement has occurred since Baja California separated from the main­land; earlier movement may have consisted of sliding along the coast, as along the San Andreas Fault today.

To summarize the plate tectonics, the land in southern California and Baja California has had a remarkable geological history: it has been car­ried successively in three different directions as part of three different geological plates. First, at various times before thirty million years ago this land consisted of scattered pieces carried to the east on the Farallon Plate. Then, subduction of the Farallon Plate scraped these bits off and as­sembled them as a scramble of suspect terranes against the leading edge of the overriding North American Plate, and their movement reversed to westerly. Finally, beginning about 27 million years ago, the same land began to be shoved across the East Pacific Rise onto the Pacific Plate, where it is now moving mostly to the north.

Scenario for the Evolution of Dudleya

And what does all of this have to do with Dudleya? In short, this land provides the habitats for Dudleya, although it is not clear how long the genus has been here and how much of this geological history it has experienced. The species of Dudleya occur primarily on rocks and cliffs near the coast in this area. Probably they originated somewhere on these rocks, perhaps when they lay farther south. The climate here is strongly moderated by the sea and less subject to the prolonged extremes of cold and drought that have characterized the interior of the continent, and it has probably not changed very' much during these geological times. The jumble of suspect terranes provides many different kinds of rock and substrates, and the cliffs of these materials face in different directions and provide many differ­ent sorts of habitat. These variables doubtless contribute to the very local distribution and endemism of many dudleyas.

To speculate a bit, this history suggests the following scenario for the evolution of Dudleya. At least by 5.5 million years ago, possibly long before, Dudleya, or its unique progenitors, had already diverged from its putative common an­cestor with Echeveria, which is probably its clos­est relative. The genus had become adapted to life on cliffs and rocks of the mixed suspect terranes along and near the coast, mostly at rela­tively low elevations and perhaps somewhat south of its present distribution. Probably Echeveria was then, as it is now, a genus adapting to higher elevations in the interior, away from the coast. Dudleya or its progenitors must already have been aboard 5.5 million years ago, when the Baja Cal­ifornia peninsula separated from the mainland of Mexico, completing its geographic isolation from old relatives there. The present habitats of Dudleya farther north had earlier been pushed from the North American Plate onto the Pacific Plate and were already sliding northwestward, and there is at least a slight possibility that Dudleya's progenitors were already there 27 million years ago when this movement began. Axelrod and Cota (1993) have recently proposed a rather similar scenario to account for the distribution of the Monterey pine (Pinus radiata) and its fos­sils and relatives.

During the last two million years, the several advances and retreats of Pleistocene glaciers to the north and on the high mountains were ac­companied by cycles of drastic change in tem­perature and rainfall at localities in Mexico away from the coast. These changes forced major mi­grations of the flora, south and back north and down and back up the mountain slopes. Envi­ronmental stress was severe, and this doubtless accelerated the rate of evolution of species on the Mexican mainland, caused many extinctions, and resulted in a more diverse array of species and genera there. However, the ocean strongly tempered conditions in the coastal habitats of Dudleya. The climate there was much more sta­ble, and any alternating periods of hot and dry and wet and cold were much less severe than those endured by its old relatives on the main­land. Thus pressures on Dudleya for change were weaker, and its evolution likely proceeded at a slower pace than in its mainland relatives. This implies that Dudleya may still bear a closer re­semblance to their hypothetical common ances­tor.

Except for the few recent immigrants in both directions mentioned above, Dudleya has evolved in total isolation from its closest relatives for at least these 5.5 million years and perhaps for con­siderably longer. This view is supported by the apparent difficulty or inability of Dudleya to hy­bridize with any related genera. Could these dif­ferences in the histories of their habitats also be related to the cytological differences between Dudleya and Echeveria in the patterns of evo­lutionary change in their chromosomes men­tioned above?

I welcome any comments that these ideas may provoke or incite.

Acknowledgments

I thank Dr. Reid Moran for many helpful suggestions and for the photo of authentic Dudleya pulverulenta and Bente Starcke King for preparing the map.

References

Atwater. T. 1989. Plate tectonic history of the north­east Pacific and western North America. Pages 21-72 in The Geology of North America. Vol. N: the eastern Pacific Ocean and Hawaii. (See also papers by J. E. Spencer and W. R Normark and by P. Lonsdale in the same volume.)

Axelrod. D. I., and J. Cota. 1993. A further contri­bution to closed-cone pine (Oocarpae) history. American Journ. Bot. 80: 743-751.

Baker, J. G. 1869. Refugium Botanicum: Tab. 66.

________. 1870. Refugium Botanicum: Tab. 200.

Berger, A. 1930. Crassulaceae. In A. Engler, Die natürlichen Pflanzenfamilien. 2nd ed. 18a: 352-485.

Britton, N. L., and J. N. Rose. 1903. Bull. New York Botanical Garden 3: 12.

Deleuil J. B. A. 1875. La Belgique Horticole 25: 371.

Jacobsen. H.. and G. Rowley. 1958. Some name changes in succulent plants. Nat. Cact. Succ. J. 13: 75.

Kearney. T. H., and R. H. Peebles. 1951. Arizona flora. Univ. of Calif. Press, Berkeley.

Moran, R. 1951. Natural hybrids between Dudleya and Hasseanthus. Bull. S. Calif. Acad. Sci. 50: 57-67.

________. 1969. Sedum in Baja California. Cact. Succ. J. 41: 20-25.

Morren. E. 1874. Esquisse du genre Echeveria DC. La Belgique Horticole 24: 153-170.

________. 1877. La Belgique Horticole 27: 248-249.

Redfern, R. 1983. The making of a continent. Times Books. New York.

Sullivan, W. 1984. Landprints. Times Books. New York.

Uhl, C. H. 1977. Cytogeography of Sedum lanceolatum and its relatives. Rhodora 79: 95-114.

________. 1985. Chromosomes of Mexican Sedum: V. Section Sedum and subgenus Sulcus. Rhodora 87: 381-423.

________. 1992a. The San Andreas Fault plate bound­ary and a natural boundary within the genera of the American Crassulaceae. The Sedum Society Newsletter No. 21: 13-15.

________. 1992b. Polyploidy. dysploidy. and chromo­some pairing in Echeveria (Crassulaceae) and its hybrids. American Journ. Bot. 79: 556-566.

________. and R. Moran. 1953. The cytotaxonomy of Dudleya and Hasseanthus. American Journ. Bot. 40: 492-502.

Van Houtte, L. 1873. Flore des serres et des jardines de l'Europe. Plate 1927-1928.

Walther, E. 1972. Echeveria. Calif. Acad. Sci.. San Francisco.

© Cactus & Succulent Journal of America, 1994