The Monthly Fern—More Quirks of Quillworts

Jon Keeley with several of his beloved quillworts (date unknown).

Once again The Monthly Fern series is featuring the Prairie Quillwort and its relatives—genus Isoetes. One post was not enough for these fascinating plants! Not only are they the sole survivors of a plant group that dominated 300 million years ago (see last month's post), they use CAM photosynthesis (1) ... that's astonishing! In fact it's so unexpected that when Jon Keeley announced it 40+ years ago, he was written off as ignorant (Keeley 2014).

Here's the conundrum. CAM photosynthesis is thought to have evolved in flowering plants (angiosperms) in hot arid environments. Many succulents, including most cacti, are CAM plants. But quillworts are primitive spore-producing lycophytes predating flowering plants by c. 200 million years. And almost all are aquatic.

Isoetes and other lycophytes split from ferns and seed plants long ago (source; black, red labels added).

Lycophyte diversity by Kingfiser (click link for full names and more info).

When I was an undergraduate long ago, only one type of photosynthesis was known (or so we were taught). As a grad student a decade later, I learned there were three: C3 is the common type; C4 and CAM are restricted to certain groups and situations (2). Since then I've largely ignored photosynthesis. But when I read that quillworts are CAM plants, I was intrigued! It was time to learn more. (Information here is from Khan Academy's Biology Unit 8, Photosynthesis unless noted otherwise.)

Photosynthesis: 1st stage powered by sunlight; 2nd makes food & oxygen for us to consume.

Photosynthesis is complicated and very chemical, but the basic process is simple. There are two stages. In the first, energy from sunlight is captured and converted to chemical energy. In the second, this chemical energy is used to convert carbon dioxide and water into glucose and similar carbon-based compounds, releasing oxygen in the process.

These are the benefits we reap. We consume carbon-based compounds for energy and to build proteins, DNA, muscles and more. And we breathe oxygen. If photosynthesis were to stop, we would die—either starve or suffocate.

As wonderful as photosynthesis is, there's room for improvement. The widespread C3 type, used by 85% of plants, is surprisingly inefficient. Carbon dioxide is captured and a sugar molecule created only about 65–80% of the time. The problem lies with an important but indiscriminate enzyme—rubisco—which will happily bind oxygen instead of carbon dioxide if given the chance (more here). 

Rubisco is the "molecular equivalent of a good friend with a bad habit" (KA, modified slightly).

This inefficiency is significantly less in C4 and CAM photosynthesis. But there's another problem and it's a big one—water loss. Plants take in carbon dioxide from air via stomata (pores) on leaf surfaces. But water vapor is lost at the same time, especially on hot dry days. Many plants close their stomata at night to prevent water loss, and when it's hot, some close stomata during the day as well. But then there's no source of carbon.

This is where CAM plants excel. They open their stomata at night and collect carbon dioxide, storing it for use the next day when the sun is shining. That way they can photosynthesize without opening their stomata and losing water. So clever!!

CAM photosynthesis requires extra energy compared to the common C3 type, but apparently it's worth the cost. CAM is used by at least 16,000 species, c. 7% of all plants. Most are desert plants, including at least 99% of the 1700 species of cacti (source). And then there are the quillworts, nearly all of which are aquatic at least part of their lives. Why would they bother with energy-expensive CAM?

Isoetes melanopoda, Prairie Quillwort, uses CAM even though it's aquatic (©2015 Robbin Moran).

Prairie Quillworts photosynthesizing by the light of day, with CO2 they gathered before dawn (Andrey Zharkikh).

Like terrestrial CAM plants, aquatic quillworts gather and store carbon dioxide at night but for a different reason. Terrestrial CAM plants have no access to CO2 during the day because their stomata are closed to prevent water loss. Quillworts have no risk of water loss, but for them daytime uptake of CO2 is difficult. It diffuses poorly in water to begin with, and most of the other plants in the pond are better at sucking it up for photosynthesis (4).

By the end of the day, the amount of CO2 in pond water is quite low. But as soon as night falls and photosynthesis stops, it quickly rises. "This must be when quillworts open their stomata to collect CO2" you may be thinking—as I did. But then a memory floated to the surface. Stomata don't work underwater! Aquatic quillworts have none, or non-functional ones at most.

From what I've read, there's still much to be learned about carbon dioxide uptake in Isoetes. However we do know that it varies with species and habitat. The few quillworts that are fully terrestrial—never submerged in water—have functional stomata and use C3 photosynthesis. They never use CAM, nor can they be converted to CAM even by keeping them underwater for a long time.

Isoetes histrix, Land Quillwort, is terrestrial (but often reported as aquatic). Late season photo by Sam Thomas; added insert by Peter de Lange.

Those quillworts that live part of their lives submerged, for example in vernal pools, are impressively versatile. They utilize CAM until water is low enough to expose their leaf tips to air. Then the stomata start to become functional and C3 photosynthesis begins to take over, progressing down each leaf cell by cell keeping just above the water! (Keeley 2014)

Isoetes howellii in dried vernal pool. It was in Howell's Quillwort that Jon Keeley stumbled upon CAM photosynthesis. © 2004 Carol W. Witham.

The many quillworts that are entirely aquatic are more puzzling. They have no stomata and their leaves are covered with a waxy cuticle. And yet they thrive, especially where other plants can't.

Aquatic Isoetes lacustris, the Lake Quillwort (Alina Ambrosova).

Isoetes lacustris in its favorite environment—lake bottom with sparse vegetation (5). (Alina Ambrosova) 

Aquatic quillworts seem to be more common in oligotrophic waters, where nutrients are scarce and there's little competing vegetation. So how do they survive if other plants can't? Probably with their unusual roots (6).

These roots have a large central air cavity that accumulates carbon dioxide gathered from sediments. Next to the cavity is bundle of vascular tissue that delivers it to the plant above. Furthermore, being CAM plants they collect CO2 at night as well as during the day, thereby doubling their harvest. Sometimes they truly flourish, covering the lake bottom in a dense green underwater carpet! (Moran 2004)

And with that, I will close. As you may suspect, this was one of my more challenging posts. Just when I had everything figured out, another puzzle would present itself. But I'm not complaining. In fact that's what I enjoy most about getting to know plants—pondering and unraveling their many little mysteries. And I know that the next time I meet up with a quillwort, it will be far richer experience.

So primitive, so simple in form, and yet so alluring (Isoetes englemannii, Nathan Aaron).

Notes

(1) C2 carbon concentration is sometimes considered a type of photosyntheses.

(2) CAM refers to crassulacean acid metabolism. To be clear, there is no "crassulacean acid"; the name refers to acid metabolism in the family Crassulaceae, where CAM was discovered (source).

(3) The widespread occurrence of CAM likely is due to repeated convergent evolution. After sequencing the pineapple genome, Ming et al. (2015) concluded that CAM arose from relatively simple reconfiguration of C3 pathways. See also Wickell et al. 2021.

(4) Many aquatic plants collect CO2 via bicarbonate; it appears that quillworts are unable to do this (Keeley 2014).

(5) Is that an alga on the leaves of Isoetes lacustris? If so, it might affect light capture but not CO2 uptake, which is done by the roots.

(6) Isoetes lacustris roots look very much like the fossilized roots of Lepidodendron trees, its ancient relatives.

Sources (in addition to links in post)

Keeley, JE. 1981. Diurnal acid metabolism in vernal pool Isoetes. Madroño 28:167-171. BHL

Keeley, JE. 1998. CAM Photosynthesis in submerged aquatic plants. The Botanical Review 64:122–158. PDF.

Keeley, JE. 2014. Aquatic CAM photosynthesis: A brief history of its discovery. Aquatic Botany 118: 38–44. http://dx.doi.org/10.1016/j.aquabot.2014.05.010

Lane, N. 2010. Life Ascending: The Ten Great Inventions of Evolution. WW Norton & Co.

Moran, Robbin. 2004. "Some Quirks of Quillworts" in A Natural History of Ferns. Timber Press.

Wickell, D, et al. 2021. Underwater CAM photosynthesis elucidated by Isoetes genome. Nat Commun. 12:6348 (open access).

Monthly Ferns—Prairie Quillwort & Scale Tree

Prairie Quillwort is 50 centimeters tall (© 2015 Robbin Moran); Scale Tree is 50 meters tall (source). Both are lycophytes—formerly "Fern Allies".

This episode of The Monthly Fern was going to feature Isoetes melanopoda, the Prairie Quillwort, mentioned last month in the wildly popular Prairie Spikemoss post (1). But in my search for information I fell down a rabbit hole and landed with a splash in an ancient Wonderland—a wet lush forest 350 million years ago near the start of the Carboniferous Period (2). After extricating myself from the muck I looked around. Tree ferns, horsetails and dragonflies looked familiar, though a bit large. But the trees were very strange.

It was during the Carboniferous that wetland forests with tall trees first appeared in the fossil record. These were hot humid riotous tangles of vegetation growing in shallow water and muddy peat that reeked of decay. Dense stands of curious trees rose high above the understory. The most common (or best preserved) was Lepidodendron, the Scale Tree. "Scale" refers to the distinctive bark—a network of diamond-shaped leaf scars (Halliday 2022).

Wetland forest of the Carboniferous; large Lepidodendron on left (Meyers Konversationslexikon 1885–1890).

Though they've been gone for 300 million years, we know a lot about these trees. Their fossilized remains are among the most extensive for any plant from any geological period, and for good reason. Not only was Lepidodendron large and ecologically dominant, it lived in waterlogged conditions conducive to preservation. Paleontologists have been able to describe features ranging from spores to leaves to trees, and even stands of trees (Hetherington et al. 2016).

Lepidodendron differed in many ways from the trees we know. Stems of young trees were covered in long ascending needle-like leaves. These fell off as the tree grew taller and wider, leaving a spiraling network of diamond-shaped scars on the trunk. With age the trunk developed a thick tough outer layer—bark of sorts—but underneath was soft spongy tissue instead of wood. At maturity the stem branched dichotomously (repeatedly forked), forming a high crown as much as 50 meters above the ground. The final branchlets were tipped with strobili (cones) filled with spores, to be dispersed by wind (source). This so-called "tree" was an arborescent lycophyte, a fern relative.

Juvenile and mature Lepidodendron on left; trees to right are related lycophytes. Reconstruction from fossils, by Falconaumanni.

Given Lepidodendron's massive build, its lack of internal wood and the waterlogged habitat, it seems it would fall before reaching such heights. What kept it upright? Some credit the thick tough bark. But others argue convincingly for the robust root system. From very long rhizomes grew a profusion of highly-branched rootlets covered in root hairs—on the order of 26,000 rootlets per meter of rhizome! They intertwined with those of adjacent trees, forming a strong anchoring network—"trees holding onto one another for stability" (Hetherington et al. 2016; Halliday 2022).

Lepidodendron truly was an arboreal superstar, dominating the wetland forests and producing immense amounts of biomass for tens of millions of years (3). But it was doomed. By 300 million years ago, the Carboniferous rainforests and arborescent lycophytes were gone, destroyed by widespread drought. Lepidodendron's only surviving relatives are little herbaceous plants—Isoetes, the quillworts.

Bolander's Quillworts in a lake in the Wasatch Mountains, Utah. Andrey Zharkikh photo.

Now we return to the present—to a roadside waterhole along highway #44 in Mellette County, South Dakota, where W. H. Over (4) made his 15,878-th plant collection on July 10, 1924. Isoetes has not been collected in the state since.

Over correctly identified his collection to genus—Isoetes. Four months later, TC Palmer called it I. melanopoda, and Daniel F. Brunton agreed in 1995. The specimen resides at the Academy of Natural Sciences, where it has been digitized for all to enjoy (cropped here).

Like all quillworts, the Prairie Quillwort has no stem. Instead a cluster of leaves develops from the rhizome. From a distance these look like clumps of grass, but up close the leaves are distinctive—long, slender, quill-like, and bright green. The leaf bases are broad and pale, forming a swollen rootstock. These usually become black with age, hence the alternative common name—Black-foot Quillwort (plants with leaf bases that remain pale have been called I. melanopoda f. pallida).

Isoetes melanopoda; MWI photo.

Being a lycophyte, Prairie Quillwort produces no seeds. Nor does it have rows of spore-bearing sori on the undersides of its leaves as do most true ferns. Instead, spores are produced in a sac on the inner side of leaf bases.

Isoetes melanopoda; MWI photo.

Surely some readers are wondering ... how is little Isoetes like immense Lepidodendron? Why do botanists think they're related? Answer: It's the roots, especially the way they branch.

"this architecture is conserved among [Lepidodendron's] only extant relatives, herbaceous plants in the Isoetes genus. Therefore, despite the difference in stature and the time that has elapsed, we conclude that both ... have the same rootlet system architecture." (Hetherington et al. 2016)

Lepidodendron and Isoetes rootlets branch dichotomously, narrowing in a stepwise manner. A is a diagram of a rootlet with 4 levels of branching. B and C show rootlets of Isoetes and Lepidodendron (scale bars are 5 mm). Hetherington et al. 2015, Fig. 1 in part.

A final question: Is Prairie Quillwort gone from South Dakota? Has it suffered its own extinction? Maybe so. Skilled botanists have searched for it with no luck. But perhaps they were limiting themselves, looking only in "roadside waterholes" and such. Of great interest to me is its occurrence nearby in Minnesota where it's a state Endangered species. It's distribution there is quite limited—rainwater and seepage pools in quartzite rock outcrops, in the southwest corner of the state.

MWI photo.

This rock—Sioux quartzite—also crops out in the southeast corner of South Dakota. I visited several locations a few years ago, and am tempted to go back now that I have a such a vivid search image for Prairie Quillwort in my head!

Sign at Palisades State Park, SD. Pink marks exposures of Sioux Quartzite in southwest MN and southeast SD.

See any habitat?

Notes

(1) Last week I noticed that views of my Prairie Spikemoss post had skyrocketed. Now Desert Mountain is getting the same level of attention. I'm suspicious. Are chatbots visiting, searching for information? Are AI models being built? Have you had this experience?

(2) In North America the Carboniferous often is treated as two periods— Mississippian followed by Pennsylvanian.

(3) Though long extinct, Lepidodendron remains vitally important. Fossilized remains of those wetland forests with giant trees and abundant peat (aka Coal Forests) drove rapid industrialization in the 18th and 19th centuries, and continue to sustain our dependency on fossil fuels (oil and gas are more common in Cretaceous rocks).

(4) W. H. Over must have been a bright and highly motivated autodidact. He quit school in Illinois after finishing the 8th grade, homesteaded in South Dakota, and by the time of his Isoetes collection, was Museum Curator at the University of South Dakota. He would go on to become one of the state's great botanists. Among his many achievements is the "Flora of South Dakota"—the first comprehensive treatment of plants known for the state. For more about Dr. Over (he was awarded a Doctor of Science degree at age 70), start here and here. I'm still looking for a comprehensive biography.

Sources (in addition to links in post)

Halliday, T. 2022. Otherlands; a Journey through Earth's Extinct Worlds. Random House.

The Carboniferous Period is in Chapter 11, "Fuel". Halliday's descriptions of past worlds are surprisingly detailed, and supported with many citations.

Hetherington, AJ, Berry, CM, Dola, L. 2016. Networks of highly branched stigmarian rootlets developed on the first giant trees. PNAS 113:6695–6700.  https://www.pnas.org/doi/full/10.1073/pnas.1514427113

Moran, Robbin. 2004. A Natural History of Ferns. Timber Press.

Desert Mountain—Utah's latest GeoSight

"now a quiet, remote monument to that violent geological time." (Smith 2025)

Desert Mountain's peaks and ridges lower right quarter of photo; white mark is high point (Google Earth).

Desert Mountain is small mountain—an isolated cluster of low peaks, ridges and knobs in Utah's West Desert. The first geologist to write about it thought "Desert Hills" more appropriate (Loughlin 1920). But its story is huge—complicated and filled with drama. And being geological, it's long.

We could start 300 million years ago, when collisions on both sides of a young North America were deforming it far inland, for example in today's western Utah. Or we could start 300 million years before that, when western Utah was covered in shallow water of the great Paleozoic Sea. Or we could go back yet another 300 million years to the creation of that sea, when the supercontinent Rodinia was coming apart. But we won't. Instead we'll start three months ago, on a hot spring day.

Shortly before I left home, the spring issue of Survey Notes showed up in my mailbox. Inside was a GeoSight—a new one, and in the general area of my travels. Of course I would go there! Of the many resources offered by the Utah Geological Survey (UGS), my favorite is GeoSights. I visited my first, the Honeycombs, in 2012. Awed by the rocks and their story, I've been geotripping to GeoSights ever since.

Sunset on the Honeycombs, 2012.

My visit to Desert Mountain was about 80 miles round trip from Delta. I took UT Hwy 6 north to the Jericho Callao Road, then drove west. Pavement soon gave way to gravel, a bit rough in places but generally good. The road crossed open juniper woodlands and sparse dry grasslands, with expansive playas to the south. After 22.5 miles, with Desert Mountain visible nearby, I stayed left at a junction and was soon at its base.

Approaching Desert Mountain from the north. Kelly Hewitt photo via Google Earth.

Rocks abound, trees not so much; pale band close to road is a fence covered in tumbleweeds.

The road continued along the base of a steep slope with granite outcrops, then climbed a short distance to Desert Mountain Pass where there was no shade to be had. I parked and reread Jackson Smith's GeoSight article inside the van, cooled by light breezes wafting through opened windows and doors.

Desert Mountain was born c. 30 to 40 million years ago, during the Great Ignimbrite Flareup which ravaged much of Nevada and western Utah. For 15 million years large volcanoes, supervolcanoes and complexes of supervolcanoes (1) produced on the order of 5.5 million km3 of volcanic material—great clouds of ash that blocked the sun, huge volumes of rock fragments hurled hundreds of miles, and searing pyroclastic flows that destroyed everything in their path. For comparison, the 1980 Mount St. Helens eruption produced only one km3 of material (source).

Geologic map shows rock units discussed here. Blue B's mark shoreline of glacial Lake Bonneville, black lines are faults, labeled arrows are mine (Smith 2025).

Today's rock outcrops suggest that the life of the Desert Mountain volcano had three stages. In the first, viscous rhyolite oozing from vents formed thick deposits of lava. Rhyolite outcrops are the remains of this relatively peaceful eruption (not part of my tour).

But while the lava oozed ever so slowly, trouble was brewing below. Gas was accumulating in the viscous magma, increasing in pressure until it literally exploded. Massive amounts of rock fragments and ash were sent flying. These pyroclastic deposits are said to be common east of the mountain, a project for a cooler day.

The eruption largely emptied the magma chamber, causing the roof of the volcano to collapse. The result was a caldera—a very large bowl-shaped depression. When the roof collapsed it broke up into a mishmash of preexisting rocks and erupted material, forming today's volcanic breccia. I may have seen it at Desert Mountain Pass, adjacent to the beautiful pale granite outcrops.

A close look at the volcanic breccia of Desert Mountain (UGS).

Desert Mountain Pass. Is that volcanic breccia behind the granite? I thought so at the time.

From the pass I drove south, using the geologic map to figure out what I was seeing. Occasionally I spotted large outcrops of much darker rock. This is granodiorite, an older intrusion predating the Desert Mountain volcano (age unknown).

Utah Juniper on pale granite; dark granodiorite in distance.

The star of the show, hands down, was the beautiful pale granite, the youngest and most extensive of Desert Mountain's rock outcrops. It was not present during the cataclysmic eruption, arriving later in the volcano's life. In the third stage, remaining magma rose but didn't reach the surface. Instead it cooled deep enough to form visibly crystalline rock—an exceptionally pale granite sometimes called leucogranite (2).

That may have been the final eruption but obviously there's more to the story, for the granite no longer is fully buried. Exhumation started about 17 million years ago, when the part of North America between the Wasatch Mountains and the Sierra Nevada (today's Basin and Range Province) began to stretch east to west. This extension deformed and fragmented older landscapes, including Desert Mountain. The caldera was uplifted, tilted and fractured, allowing erosion to slowly expose and sculpt the lovely granite.

Desert Mountain granite. Sonny Wilson photo, via Google Earth (cropped).

Granite on the west side of Desert Mountain. gjagiels photo, via Summit Post.

Spectacular outcrop at south end of Desert Mountain. Hmmm ... what are those black and white bands?

There's one more chapter in the Desert Mountain story. It's incomplete, difficult to properly place in the overall timeline, and has rock classification issues. But there's also fun to had.

Sometime after the granite was intruded—perhaps while the magma was cooling or later during extension (or both)—molten material filled fractures forming dikes. Whitish aplite dikes formed first, followed by dark andesitic dikes (3). How do we know the order? By their cross-cutting relationships! These are fun to find and worthy of attention for the story they tell. At Desert Mountain, the white dikes cross the pale granite and are therefore younger. The dark dikes cross both the white dikes and the granite and are therefore the youngest of the three.

 Wonderful display of cross-cutting relationships; white arrows mark less conspicuous aplite.

I considered camping at the base of this outcrop but it was much too hot for me. On the drive to Delta, I stopped and took one more photo of the beautiful pale granite. Then I continued south.

See the dike?

Notes

(1) Desert Mountain is part of the Thomas-Keg-Desert mountains caldera complex (DeCourten 2003). The Honeycombs, mentioned early in the post, may be related.

(2) The pale granite at Desert Mountain was called leucogranite early on (e.g., Kattelman 1968). Now "leucogranite" is increasingly used for a pale granite formed in collisional tectonic settings, for example in the Himalayas (E. H. Christiansen, personal communication). For more, see Miller's excellent Perspective (2024). He explains that though collisional is by far the most common tectonic setting for leucogranite formation, it can form in others, including extensional, if certain conditions are met (e.g., composition low in aluminum). In any case, "leucogranite" is used in Jackson's GeoSights article; Christiansen prefers "granite". [Suggestion to UGS: In GeoSights articles, cite a few sources for additional information.]

(3) The dark dikes also are controversial. According to Jackson, the rock "apparently" is very dark lamprophyre—a catch-all term for various peculiar ultramafic rocks not amenable to the usual classifications (source). Christiansen and colleagues prefer andesite.

Sources (in addition to links in post)

I'm grateful to Eric Christiansen, Professor Emeritus at Brigham Young University, for answering my questions about igneous rocks at Desert Mountain, and for his appreciation of cross-cutting relationships.

Brigham Young University. 2013. Supervolcano in Utah: massive ancient volcano discovered by BYU geologists. YouTube.

DeCourten, FL. 2003. The Broken Land; adventures in Great Basin geology. U. Utah Press.

Loughlin, G. F., 1920, Desert Mountain, in B. S. Butler, and others, Ore deposits of Utah: U. S. Geol. Survey Prof. Paper III, 444-445. PDF

Miller, CF. 2024. Granites, leucogranites, Himalayan leucogranites ... Elements 20(6):359–364. doi: https://doi.org/10.2138/gselements.20.6.359

Rees, DC, Erickson, MP, Whelan, JA. 1973. Geology and diatremes of Desert Mountain, Utah. Utah Geological & Mineralogical Survey Special Studies 42. PDF

Smith, J. 2025. Geosights: Desert Mountain, Juab County, Utah. Utah Geological Survey, Survey Notes.

The Monthly Fern??—Prairie Spikemoss

Selaginella densa—moss, fern, fern ally, or none of the above? Coin is 19 mm across.

This month the South Dakota fern series features another oddity—a spikemoss, genus Selaginella. It's even more unusual than last month's Water Clover, for while water clovers are ferns, spikemosses are not, at least not anymore. So where in the greater scheme of plant classification do they belong?

The Prairie or Dense Spikemoss, Selaginella densa, is the more common of South Dakota's two spikemosses. It occurs in the Black Hills and scattered across the west half of the state. If you live in or have wandered across the Great Plains or Rocky Mountains, you may have seen it, for it grows on a wide range of sites—prairies, alpine meadows, dry rocky slopes, rock crevices, sandstone, quartzite or granite rock, and dry gravelly, clayey or sandy soil (Flora North America). Or maybe you overlooked it, as I used to do. After all, it looks very much like a moss (1).

Spikemosses were first classified—given a name and assigned to a plant group—by the great Swedish botanist Carl Linnaeus, founder of today's system of naming organisms. In his Species plantarum (1754) he placed them in the CRYPTOGAMIA MUSCI section—the mosses. That was a big mistake, but at the time it was a reasonable decision. Like mosses, Selaginella produces spores (2). But unlike mosses, it has vascular tissue—plumbing for transporting water and nutrients.

In the late 1890s another Swedish botanist was studying spikemosses, while preparing a Catalogue of the Flora of Montana and the Yellowstone National Park. Per Axel Rydberg had emigrated to the United States in 1882, hoping for a career as a mining engineer. But after a serious injury in an iron mine in Michigan, he moved to eastern Nebraska to teach mathematics. He also studied botany at the University of Nebraska—the beginning of a "lifelong devotion to plant studies" in the Great Plains and Rocky Mountains (source).

In 1895 and 1896, Rydberg was sent to Montana by the US Department of Agriculture to collect grasses and forage plants. The next summer he returned, with the first field expedition of the New York Botanical Garden. He made about 1800 collections representing 800 species—20,000 specimens in all (replicates were collected for exchange or sale to other institutions).

Rydberg's Catalogue included a large foldout map showing localities mentioned in the text. He noted that the eastern half of the state was "practically unexplored botanically." (BHL)

Going through his collections that winter, Rydberg saw that the flora of Montana was poorly known, even with his additions. "It was therefore considered advisable to extend the work and study all the material from the state that was accessible." He examined specimens from 16 institutions and private individuals, ranging from the Lewis & Clark collection (1803–1804) to the Montana Ladies' [Columbian] World's Fair Set (1893). By the time the Catalogue was published in 1900, Rydberg had added 776 species to the flora of the Rocky Mountain region, including 163 novelties—species new to science (Rydberg would become known as a notorious splitter).

Among the novelties was a low densely-tufted plant with very short stems covered in bristle-tipped leaves 3–5 mm long. Fertile stems were taller, to c. 4 cm, with spore-bearing leaves (sporophylls) neatly arranged in four ranks, forming terminal strobili (aka cones).

Prairie Spikemoss forms dense mats in this soil crust. Matt Lavin photo.

Selaginella densa's 4-angled strobili rise above very short sterile stems that look like clusters of bristle-tipped leaves. cinthyadasilva photo.

Rydberg knew the plant was a spikemoss, but the dense "moss-like" form was not something he had seen before. After careful study of seven specimens, he concluded it was a new species, calling it Selaginella densa. The holotype (basis for formal description) was a specimen collected in 1889 by Valery Havard, a French-born American military physician, explorer and botanist.

Havard identified his specimen (NYBG) as S. rupestris, which is widespread in the east half of the US.

In his Catalogue Rydberg followed the accepted classification of the day. He included Selaginella densa in the Pteridophytes—spore-bearing vascular plants, mainly ferns. He put it near the end of the section, with horsetails, clubmosses and other oddballs. These were the Fern Allies. Like ferns they bore spores, had vascular tissue, and reproduced via two distinct independent life stages. But otherwise they were decidedly unfernlike, and very different from each other. Just look below!

Horsetails and scouring rushes, genus Equisteum, have jointed stems with cylindrical sheaths tipped with teeth. These are thought to be highly modified leaves. Spores are born in terminal cones.

Unbranched species of Equisetum are called scouring rushes. Andre Zharkikh photo.

Whisk ferns, genus Psilotum, have linear shoots that fork in the upper half. Minuscule scale-like leaves subtend globose spore containers 2–3 mm across.

Sideways view of a whisk fern. Mary Keim photo.

Quillworts, genus Isoetes, are aquatic, with grass-like clusters of linear leaves. Spores are born in sac-like structures in enlarged leaf bases.

Bolander's Quillworts in a lake in the Wasatch Mountains, Utah. Andrey Zharkikh photo.

Clubmosses, family Lycopodiacee, are a more diverse group, with 7 genera and 27 species in North America. Some are suggestive of spikemosses; in fact spikemosses were put in the genus Lycopodium by Linnaeus.

These clubmosses, all formerly genus Lycopodium, are now 4 separate genera; from Ferns and Evergreens of New England, 1895 (BHL).

The Fern Allies group came into use in the early 1800s, as a catchall for diverse, puzzling, somewhat fernlike plants. But after about a century botanical experts began to object. Some Allies appeared to be more closely related to ferns, others not so much. Then less than a century later, the Allies got caught up in a revolution. Biologists were switching to a phylogenetic approach to classification. In a nutshell (a very tiny one), they now hope to classify organisms based on evolutionary relationships, i.e., so that all members of a group share a common ancestor. The Fern Allies do not, so they were reclassified (3).

The commonly accepted classification splits the Allies into two groups that diverged long ago, early in the evolution of vascular plants. One includes ferns, horsetails, whisk ferns and seed plants. The other group is much smaller, a collection of relatively primitive plants: quillworts, clubmosses and spikemosses. These are lycophytes (answer to question at top of post). For a longer summary, see The Ferns and their Allies at Cliffnotes. For a deep discussion, start with Pteridophyte taxonomy on Wikipedia.

Fern and lycophyte classification from the Pteridophyte Phylogeny Group. Black labels added, not sure how that guy in the corner snuck in.

Notes

(1) Not all spikemosses are as humble and mosslike as ours. Selaginella is a large genus with c. 800 species, mainly of the tropics and subtropics. In hospitable habitat, spikemosses can be quite showy—some are iridescent!

Selaginella uncinata, Blue Spikemoss, is native to moist shady sites in southern Chile and is widely cultivated; leaves are 3–4 mm long (Flora North America). GKA Dickson photo.

(2) Actually Linnaeus couldn't decide whether "fern dust" was pollen or seeds. The concept of spores would come later.

(3) It's been really hard to give up Fern Allies! It's such a handy label for those diverse kinda-fernlike species. Not surprisingly, the name hasn't gone away. Sometimes it appears under an alias, for example "Fern Relatives" in Ferns of Northeastern and Central North America (2005). More often it pops up in casual conversation, or is used by older botanists who haven't bothered to learn the new scheme. After resorting to "Fern Allies" in a message to pteridologist Robbin Moran, I committed to learning it (Robbin is much too kind to disapprove directly, but he did refer to "lycophytes" in his reply).

Sources (in addition to links in post)

Cobb, B, Farnsworth, E, Lowe, C. 2005. Ferns of Northeastern and Central North America. 2nd ed. Peterson Field Guide Series.

Linné, Cv. 1754. Species plantarum v2. BHL

Moran, Robbin. 2004. A Natural History of Ferns. Timber Press.

Rydberg, PA. 1900. Catalogue of the Flora of Montana and the Yellowstone National Park. Memoirs of the New York Botanical Garden. BHL