Stone Magnolias

Stone R. B. Magnolias Figlar extraordinary fossil site in Idaho, seventeen-million-year-old leaves look remarkably like present-day leaves-in many cases, in better condition than At an autumn’s windfalls. just walked into look at my magnolias, several of which were in full bloom. Suddenly my attention was riveted on the evening news: Dan Rather was describing the successful extraction of DNA from a seventeen- to twenty-million-year-old Magnolia leaf found in Idaho. As a certified magnoliaphile I was immediately captivated and had to learn more. My quest led me to Charles J. Smiley of the University of Idaho in Moscow, manager of the Clarkia fossil site for many years and author or co-author of several papers on it since its discovery in 1971. With his help I was able to obtain much of what had been published on Clarkia. Later, in 1991, he invited me and my wife, Anita, to visit and collect at the site. It was April of 1990 and I had my kitchen after a History of the Site The fossil beds are located about fifty miles northeast of Moscow, Idaho, in the valley of the St. Maries River near the town of Clarkia. The principal fossil find there, called the P-33 site, is located on the property of Francis Kienbaum. In September 1971, while Mr. Kienbaum was grading a portion of his land for use as a snowmobile racetrack, he noticed that the bulldozer was turning up a lot of black leaves, some even blowing around in the wind. Fortunately, it aroused his curiosity. Noticing that most of these leaves appeared to be from broadleaf trees not common in that part of Idaho, Kienbaum telephoned the University and eventually contacted Dr. Smiley. What Kienbaum had uncovered is now regarded as the best preserved Miocene plant fossil site in the world. Those black leaves were the organic remains of leaves that fell there at least seventeen million years ago. In the ensuing years Dr. Smiley and his team of researchers discovered more than 130 different plant species in the nine-meter-thick sediments, including fossil Magnolia latahensis (Berry) Brown and perhaps one or two other species of Magnolia as well as fossil equivalents of many of its present-day associates (Golenberg et al. 1990). These include tulip tree (Liriodendron), sweet gum (Liquidambar), baldcypress (Taxodium), oak (Quercus), persimmon (Diospyros), red bay (Persea), tupelo (Nyssa), beech (Fagus). Also present in the fos- several genera now confined to Asia, such as dawn redwood (Metasequoia), China fir (Cunninghamia), katsura (Cercidiphyllum), Zelkova, and princess tree (Paulownia) (Smiley and Rember 1985). Though the assemblage of plants matches the present flora of southeastern North America more closely than that of any other region, it appears that the Miocene Clarkia flora was more diverse than any existing flora of temperate North America. sil record eastern are The author in the St. Maries River valley near the Clarkia, Idaho, P-33 site. Photo by Anita Figlar. Origin of the Clarkia Flossil Beds About twenty million years ago, during early Miocene time, widespread volcanic activity was underway in the Pacific Northwest. One of the largest known lava flows overspread eastern Washington and is manifest today as the Columbia River basalts. Farther east, in the proto-St. Maries River valley, similar volcanic activity occurred; one of the lava flows suddenly dammed the valley, creating a deep, narrow lake (Smiley and Rember 1979). The lake must have been cold-bottomed and limited in oxygen with very little microbial or scavenger activity, thus favoring the preservation of any plant parts deposited from the nearby shore. The lake silted in fast, probably from airfall ash, clays, silt, and other products of erosion, perhaps filling in completely within a thousand years or less. This gentle but rapid infill- ing of sediment on the lake bottom entombed the leaves and fruits in finely laminated sediments where they remained, saturated with water. Untouched by weather or erosion, running water, glaciers, or any of the other geological processes that have modified most of the earth’s surface, they survived the past seventeen million years virtually unaltered. Everything had to go right and it did. Exploring Clarkia We arrived in the Clarkia area by way of the University of Idaho in Moscow in early July 1991. Though it was nearly midsummer, northern Idaho looked oddly spring-like as black locust trees (Robinia pseudoacaccia), widely planted as ornamentals, were still in bloom. On the hourlong drive to the fossil site, Dr. Smiley provided a detailed geological and 5 botanical narration as accompaniment to the roadside scenery. Basaltic extrusions give evi- dence of the Miocene lava flows in the roadcuts. These are often interbedded with clay-like layers known to geologists as the Latah sediments, which are sometimes fossil bearing. As we traversed the gently rolling topography, the existing flora-grand fir (Abies gmndis), douglas fir (Pseudotsuga mensiesii), western white pine (Pinus monticola), western larch (Larix occidentalis), subalpine fir (Abies lasiocarpa), quaking aspen (Populus tremuloides)-contrasted markedly with my mental image of the rich magnolia-beech-baldcypress forest that existed in Miocene time. A sign that read "Fossil Bowl," the name of Mr. Kienbaum’s racetrack, signaled that we had arrived at site P-33. Bill Rember, Yang Hong, and other researchers were taking core samples of stream and lake sediment. As Rember told us about some of his recent discoveries, including "a really big leaf" believed to be a Magnolia, possibly belonging to section Rytidospermum (a group of closely related Magnolia species often called big leaf or umbrella magnolias), I kept looking at the ground. Leaves, black leaves, mostly baldcypress, were all over the place. Within minutes I had picked up pieces of shale containing dawn redwood, sweet gum, chestnut (Castanea), and yes, even Magnolia. Dr. Smiley suggested that we could get better specimens if we chopped them fresh from the sediment. Chopping Fossils This method calls for a pulaski, a kind of pickax, to chop out small blocks of the soft shale sediment. Individual bedding surfaces are then split off with a pocket knife. Each bedding surface usually reveals one or more fossil leaf compressions, many of which still show original green or red pigmentation for a short time before turning black. Sometimes fruits and twigs are also present. This is how we explored this ancient forest: I would chop and Dr. Smiley would split and identify while our wives, Peg Smiley and Anita An unusually well preserved Miocene fossil, an immature Clarkia Magnolia fruit aggregate, next to its present-day counterpart, Magnolia grandiflora. Figlar, wrapped and labelled specimens. As we lifted the leaves from the ancient sediments, my sense of time often became confused; for all intents and purposes, these seventeen-million- year-old leaves looked like present-day leaves, better than those that fell last fall. We had time to dig pond for only an hour or so, but in that short time we collected fossil leaves of numerous Miocene species, including the Magnolia latahensis and what appears to be a second species of Magnolia, one that resembles the extant cucumber magnolia (M. acuminata) of section and in most cases, into my own Tulipastrum. Later, we visited the main fossil collection at the Tertiary Research Center of the University of Idaho. There, Dr. Smiley showed us many of the better specimens of Magnoliaceae from Clarkia. Most impressive in this collection was an exceptionally well preserved fossil of an immature Magnolia fruit aggregate. Even more remarkable is the extraordinary resemblance of this fossil fruit aggregate to that of the presentday southern magnolia (M. grandiflora) of the section Theorhodon. Close examination of the fossil showed nine tepals, approximately 250 stamens, and some 120 carpels-all of these being well within the ranges for M. grandiflora. Also impressive among Dr. Smiley’s collection were the quality and quantity of fossil tulip tree leaves and fruits. Looking for Modern Analogues Back home in New York State,I determined to questions raised at Clarkia. Did more than one Magnolia species exist in the Miocene Clarkia flora? Do these fossil Magnolias have modern analogues? Which extant species do they most resemble? The possibility of more than one Magnolia species in the Miocene Clarkia flora was raised by Smiley and Rember (1981) in the course of a stratigraphic analysis of a column of sediment 7.6 meters deep. In this study the researchers selected a one-meter-square area at the P-33 site, then painstakingly peeled off layer after layer of sediment, each usually less than one centimeter thick. All plant fossils in each layer were collected, identified, quantified, and tabulated for the entire 7.6-meter depth of the column. The resulting data suggested that different plant communities occurred at different levels of the column. Some layers of the column revealed fossil evidence of a swamp assemblage species such as baldcypress, tupelo, and Magnolia latahensis, whereas other layers reflected a slope assemblage dominated by oak, beech, and dawn redwood. In this latter assemblage Smiley and Rember found fossils of a species of Magnolia whose leaf morphology was suggestive of present-day M. acuminata. Smiley told us that unlike those of M. latahensis, fossils of this species occurred rarely and were typically found in a more fragmented, abraided condition, indicative of long-distance transport by running water. Using both close-up photographs and actual fossil material, I compared leaf structures and venation patterns of the two fossil Magnolias to each other and to numerous species of living Magnolias. While the sample size of fossil material was small, the specimens were in excellent condition and showed leaf venation down to the smallest details. Comparative pursue the analysis of these characters indicated that the two fossil species, M. latahensis and the slope magnolia, showed a closer resemblance to living M. grandiflora and M. acuminata, respectively, than to each other or to any other living Magnolia species studied. In fact, the struc- tural details and venation parameters for both fossil species were well within the ranges for each of their proposed living analogues. Smiley’s comment that the abraided leaves of the slope magnolia suggest thin, fragile deciduous leaves, further supports the affinity between the slope magnolia and living M. acuminata. In the case of M. latahensis, both the leaf stratigraphic evidence of a swamp assemblage suggest an affinity with living M. grandiflora, a conclusion that is bolstered by the extraordinary morphological similarity of the fossil fruit aggregate to those of M. grandiflora. An even more convincing case for the latahensis /grandiflora affinity structure and the would be made if it could be demonstrated that the fossil fruit aggregate did in fact belong to M. latahensis and not to the slope species. Probability suggests that since only one type of Magnolia fruit was found at Clarkia, it came from the species that produced the most leaves-that is, M. latahensis. Unfortunately, the fossil fruit did not have any leaves associated with it, making it impossible to prove that it was produced by M. latahensis. A Change in Climate Some time after sediment had filled the lake at the Clarkia site-in the middle and later parts of the Miocene epoch-the Cascade Range formed, producing a rainshadow effect Clarkia and surrounding areas. Not only did the Cascades rising to the west cause precipitation and humidity to decrease, but this climatic barrier also allowed more frequent invasions of arctic air masses from the north. These combined effects would eventually doom Magnolia latahensis and the rest of the moisture-loving Miocene Clarkia flora in favor over 7 Paleobotanical Detection: Leaf Structure of Magnolia acuminata Compared to a Miocene Fossil In contrast to conventional botany, paleobotany deal with fragmentary evidence. The fossil leaf may be nearly pristine, but the paleobotanist does not have the luxury of examining parts of the plant that would have been attached to the leaf, nor the habitat in which it grew. This presents a special problem with Magnolia since its leaves can easily be confused with those of other genera such as the tupelo and pawpaw (Asimina) whose leaves are of similar size and shape and, like Magnolia, are entire margined (that is, without lobesl. To address this issue, we studied leaf venation patterns of most of the modern equivalents of the entire-margined Miocene Clarkia genera and found that even when the leaves were similar in size and shape, most could be distinguished from Magnolia by the number of secondary veins on either side of the primary vein. Magnolias typically have ten to fourteen of these secondary veins (except for section Rytidospermum, which ordinarily have twenty or more), while the leaves of most other entire-margined genera have fewer than ten. In two genera, Nyssa and Asimma, the number of secondary veins did not differ from those in Magnoha, but other characteristics (for example, the behavior of loop-forming branches of secondary veins) were used to differentiate them requires one to (Dilcher 1974). Leaf of to M. acuminata, however, show less acute or more acute angles of divergence of the secondaries from the midrib: M. grandiflora, 60 to 70 degrees; M. campbellli and M. I extant Magnolia acuminata. J delavayi, up to 90 degrees; M. denudata, less than degrees. Further evidence of the close relationship between M. acuminata and the fossil slope magnolia was found when we compared ultimate 45 Having found a way to distinguish Magnolia from other genera using leaf evidence only, we still needed to distinguish among the various species of Magnolia. In M. acuminata the angle of divergence of the secondaries from the midrib is usually between 45 and 65 degrees, with the angle tending to be more acute near the leaf apex. Frequently, the secondaries in the basal half of the leaf are recurvate (curving down before curving back up again), especially in proximity to the midrib. These and other venation characteristics peculiar to M. acuminata were also found to be present on fossil leaf specimens of the slope magnolia. Other extant Magnolia species with leaf shapes comparable Fossil leaf of the socalled slope magnolia. venation structures such as areoles and veinlets under forty-times magnification. Areoles are usually pentagonal or quadrangular shapes formed by the lower order veins. In M acuminata these areoles are usually 0.5 to 1.0 millimeters across. Typically more than half of these areoles contain veinlets that "dead end" inside the areoles. These features are clearly visible in the fossil and appear to closely match those of extant M. acuminata. Other extant magnolias (M campbellu, M. sprengen, and M. delavayi),however, have larger (2 to 3 mm) areoles, while in M. grandiflora and in fossil M. latahensis veinlets are rarely present or are indiscernible, possibly because of the leathery leaves. 8 of the more drought-tolerant and boreal flora found today throughout interior western North America. Other Miocene fossil records, though less well preserved, indicate that many Clarkiatype floras existed throughout western North America, including the Miocene Latah flora near Spokane, Washington, and the Miocene Puente flora in the Los Angeles area (Axelrod 1939). Fossil Magnolia corrallina Chaney, from the Miocene San Pablo (Neroly) flora of west central California, is said to resemble presentday M. grandiflora just as the Clarkia latahensis does (Condit 1938; Chaney and Axelrod 1959). More recently, in a report on fossil Magnolia seeds found in Brandon lignite of the Oligocene epoch (twenty-five to forty million years ago) in west-central Vermont, Tiffney describes two fossil Magnolia species, M. septentrionalis and M. waltonii, whose seed morphologies suggest affinity to living species of section Theorhodon and section Tulipastrum, curiously similar to the affinities described here for Clarkia fossil Magnolias. The Clarkia evidence, which shows close morphological similarity in the leaves of the Miocene M. latahensis and present-day M. grandiflora as well as other section Theorhodron species, supports the view proposed by Parks and Wendel that species that persist in the same moist habitats as their Miocene counterparts tend to retain their ancestor’s morphological characteristics. In contrast, species found in more xeric habitats today generally show greater morphological divergence from their Miocene ancestors. Grounds for Speculation The fragmentary nature of paleobotanical evidence allows, even encourages, speculation. For instance, one can speculate that some twenty-five million years before the presentduring the late Oligocene/early Miocene-aa species resembling M. grandiflora was distributed continuously from what is now far western Canada southward through the Great Basin into Mexico and Central America, then back up through south-central United States into New England and perhaps beyond. Later, in middle Miocene time-fifteen million years before the present-alarge gap in the distribution might have developed in the American Southwest as that region became more arid. At the same time, mountains rising along the western coast of North and Central America would begin starving other Theorhodon populations of moisture, resulting in many disjunctions. And finally, the New England leg of the distribution would shrink southward as the Miocene and Pleistocene climates cooled. The western North American populations would vanish by the end of the Miocene. 9 Eventually, in Quarternary time-perhaps as recently as ten thousand years ago-climatic change would force another separation of the Theorhodon distribution between Texas and disjunctions in Mexico. These Central disjuncts, having been genetically isolated for so long, have evolved into separate species (Vasquez-G 1990). M. grandiflora, R. Condit, C. 1938. The San Pablo flora of west central California. In R. W. Chaney et al., the of the Columbia Plateau. Carnegie Institute of Washington, Publication 617. Contributions American which still has a sizable continuous distribution in southeastern United States, is all that remains of the original continental distribution of M. latahensis. A similar migration could be posited for the ancestor of the cucumber magnolia (M. acuminata) except that, being a more temperate species, its range may have extended farther north, perhaps originally as far as coastal Alaska during the early Miocene epoch. The fact that the closest living relative to the cucumber magnolia, M. liliflora, is found in Asia suggests that the ancestral species could still have been genetically "communicating" across the continents via the Alaska land bridge to Asia during or shortly before the early Miocene. This relatively short period of isolation, about twenty-five million years, might have resulted in the morphological similarity that exists between the Institute of Dilcher, just the amazing at Clarkia have brought us much findings closer to understanding Magnolia’s past in North America, future studies there might someday unlock still more secrets. While this manuscript was being prepared, my mother, Adelene Greenwood Figlar, passed away. She was a kind and lovmg lady and, probably because of me, developed a great fondness for magnolias. This article is dedicated to her memory. Literature Cited Axelrod, 1938. Miocene floras from the western Mohave Desert. Contnbutions to Paleobotany, 55-56. Carnegie Institute of Washington, Publication 516. D. I. Publication 476. 344: 656-658 Parks, R., and J. F. Wendel. 1990. Molecular divergence between Asian and North American species of Lmodendron (Magnoliaceae) and implications for interpretation of fossil floras. American Journal of Botany 77(10): 12431256. Smiley, C. J., and W. C. Rember. 1979. Guidebook and Roadlog to the St. Manes River (Clarkia) Area of Northern Idaho. Moscow: Idaho Bureau of Mines and Geology Information Circular 33: 1-45. Smiley, C. J., and W. C. Rember. 1981. Paleoecology of the Miocene Clarkia lake (northern Idaho) and its environs. In A. J. Borecot, J. Gray, and W. B. N. Berry (eds.), Commumties of the Past. Stroudsburg, PA: Dowden, Hutchmson, and Ross, 551-590. Smiley, C. J., and W. C. Rember. 1985. Composition of the Miocene Clarkia flora. In C. J. Smiley (ed.), Late Cenozoic History of the Pacific Northwest. San Francisco, CA: American Association for the Advancement of Science, Pacific Division, 95-112. Tiffney, B. H. 1977. Fruits and seeds of the Brandon lignite : Magnoliaceae. Botanical Journal of the Linnaean Society 75: 299-323. as Acknowledgments Paleobotany. Carnegie D. L. 1974. Approaches to the identification of angiosperm leaf remains. Botanical Review C. section Tulipastrum species, M. acuminata and M. liliflora. These scenarios seem plausible, but for now tary fossil evidence. But to Washington, 40(1): 1-157. Golenberg, E. M., D. E. Giannasi, M. T. Clegg, C. J. Smiley, M. Durbin, D. Henderson, and G. Zurawski. 1990. Chloroplast DNA sequence from a Miocene Magnolia species. Nature two extant they are largely conjectural, given the fragmen- W., and D. I. Axelrod. 1959. Miocene Floras Chaney, Vazquez-G., J. A. 1990. Taxonomy of the genus Magnolia in Mexico and Central America. Madison, WI: University of Wisconsin M.S. Thesis. Dick Figlar is a past president of the Magnolia Society who grows more than thirty kinds of Magnoha in his own garden in Pomona, New York. He is developing a personal arboretum of Magnolia in the foothills of the Blue Ridge Mountains of South Carolina.