An Attempt at Prairie Reconstruction – Part 1

As a backyard prairie begins its second birthday, I anxiously await warming soil temperatures that will signal the growth of what has established and what arrives and germinates in the bare inter-spaces first. I have hope against the dreaded invasive Bermudagrass (Cynodon dactylon), as its runners have thinned severely and rooted less along the total length of the stolons in the young prairie’s germinal empire. But there are the strong rhizomes of Bermudagrass to contend with as well.  The 50+year old African turf here is very well established in some spots.  In others it has gone the way of the too-thirsty St. Augustinegrass (Stenotaphrum secundatum) – died and bequeathed its space to annual broadleaf forbs commonly associated with lawns, or where seeded, natives with a hard heavy seed that are able to gain a foothold in dead turf.

Here’s to hoping year 2 will show positive progress.

Reconstruction vs. Restoration

The two terms are placed under the broad category of “restoration”, which adds to some confusion.  Often, reconstruction and restoration are used interchangeably, but have different meanings.  A prairie reconstruction is the re-establishment of native plant communities on a site where such communities had been eradicated through farming or establishment of lawn.  A prairie restoration on the other hand, involves increasing and maintaining plant communities currently surviving in a remnant prairie site.  Remnant prairies are simply pieces of prairie that escaped the plow.  This does not imply the plant communities there are pristine and in a ‘virgin prairie’ state, however.  All remnant prairies are degraded to some degree and almost always have at least one identifiable exotic species invading their fragile and fragmented state.  In the northern Great Plains, it’s usually Smooth Brome (Bromus inermis), and in the southern Great Plains, it’s Johnsongrass (Sorghum halepense) or KR Bluestem (Bothriochloa ischaemum).

A Review of  Year 1: From Caribbean-grass Lawn to Mixedgrass Prairie

North Central TX – Western Cross Timbers

Site History

In the decades following the Dust Bowl and through the World War II years, this land was a former peach orchard. In my boyhood days, I climbed and played in the only remaining peach tree on this site from that agriculture venture. Following the abandonment of the peach orchard, the land was subdivided and my grandparents were the first to build a house in the new subdivision. This was the 1950s, and green lawns and European landscaping became the rage and the norm; St. Augustinegrass, Bermudagrass, Japansese Nandina, Japansese Honeysuckle, Chinese Privet, and Chinese Holly became the go-to plants.

Over the course of the following decades, cycles of drought and a changing climate dictated only the hardiest of the exotics would remain. Of the two exotic turf grasses seeded, St. Augustine and Bermudagrass, the latter has shown the highest persistence and highest areal coverage increase. The St. Augustine has decreased in population size, lacking the aggressive rhizomatous structure to resist drought as the Bermuda does. Even then, the Bermudagrass has begun to die out in a few areas. Those are the spots targeted for initial prairie reconstruction.

Reconstruction Timeline

September 2013

The proposed initial reconstruction site (approx. 285 sq. ft.) was sprayed with generic Glyphosate at 2% active ingredient to kill actively growing Crabgrass (Digitaria sanguinalis) and Bermudagrass (Cynodon dactylon), and Khakiweed (Alternanthera caracasana).

The soils at the site are a mix of Truce series (moderately acidic to neutral pH fine sandy loam atop claystone bedrock) and a human-made soil of fill sand, from when the site was leveled to support an above-ground pool, leftover construction materials (crushed Austin chalk limestone and residual concrete mix), and decades of lawn clippings that were deposited here.  The bulk of the clippings layer was raked up and hauled off-site.  What remained was removed with light tilling until the site was mostly bare soil.

The initial planting site, post-herbicide application. Note the green of the cool season plants, an excellent fire break.
The initial planting site post-herbicide application. Note the green of the cool season plants.

January 2014

Hand-harvested seed. From L to R: Rough cleaned and mixed, Maximillian Sunflower drying, screened and cleaned seed ready for planting.
Hand-harvested seed. From L to R: Rough cleaned and mixed, Maximillian Sunflower seed stalks drying, screened and cleaned seed ready for planting.

Hand-harvested seed from several counties was thoroughly mixed in preparation for seeding into the area. The site was raked heavily once more to remove remaining thatch, which had built up over the previous decades from the site being treated as a high-maintenance non-native lawn without bagging of clippings. The thatch was hauled off-site and dumped legally. Approximately 5 bulk gallons of rough-cleaned seed (72 species) was uniformly scattered over the site and watered-in for three consecutive days.

Commercial germplasms of native species (e.g. “Haskell” Sideoats Grama, “Lometa” Indiangrass, etc.) were also included in the seeding.  That seed was sown after the hand-harvested seed was sown, as commercial seed is cleaner than hand-harvested seed and often ends up in the bottom of a barrel or bag when mixed with trashy seed mixes.  Commercial seed should be sown before or after trashy seed is sown to prevent settling and uneven distribution within the mixture.

May 2014

A large patch of Redroot Pigweed (Amaranthus retroflexus) dominated the center of the burn unit, comprising roughly 60% of the vegetational area. The species was mostly in flower and easily pulled up, removing much of the taproots. The remains were carefully bagged and thrown away.  The presence of this introduced species indicates the land was in an old-field successional stage in between the years of peach orchard operations and later subdivision development.

The Redroot Pigweed patch. This species is probably the most common non-native Amranthus in North Central TX.
The Redroot Pigweed patch. This species is probably the most common non-native Amranthus in North Central TX.
Crow Poison

I would like to do a soil seed bank experiment at two other sites in this yard to see what other species besides the common lawn associates are present.  I am assuming very few native species are present in the seed bank until proven otherwise.  The only native forb species that has appeared from the seed bank is Crow Poison (Nothoscordum bivalve).

Greenhouse grown seedlings were planted within the early successional prairie matrix, utilizing the protecting canopies of Common Sunflower (Helianthus annuus) and other annuals. As of this writing, 63 seedlings have been planted, along with rhizome cuttings of Eastern Gamagrass (Tripsacum dactyloides) and Stiff Sunflower (Helianthus pauciflorus). An additional 100+/- seedlings were scheduled to be planted in late June or early July. Another round of larger seedlings (grown in 5” pots) were panted in September, finalizing the planting for the year.

May aspect. A portion of the initial 62 seedling installation is shown.
May aspect. A portion of the initial 62 seedling installation is shown.

July 2014

Growth of all species, including the undesirables, increased greatly due to the early summer rains. The Common Sunflowers had reached over 9ft. in height, and the Crabgrass, though largely unwanted, was realized as a potential annual nurse crop. I expect its coverage to decrease over the years, especially as more perennial species become established and root zones become occupied.  Spot spraying of exotics and other unwanted species continued as needed.

Phase 2 seedling installation.
Phase 2 seedling installation.

An additional 102 seedlings were installed on the eastern and western sides of the intial unit, expanding the perimeter. These were marked with blue pin flags to note second phase of seedling installation. The previous pink pin flags used to mark the initial seedling installation will remain in place through the end of the year to take stock of surviving and non-surviving seedlings. As of July, survival is in the 90% range.

Typical for the month of July in the Western Cross Timbers is a decrease in soil moisture as the May and June precipitation pattern effects fade away, leading into the hot and dry true summer months of July, August, and early September. The blue flagged seedlings were watered until establishment, which took longer than the seedlings installed in the first phase. However, the phase one seedlings were not watered any longer.

backyard prairie
July 5th, 2014.

In the photo above, the young prairie showed its aestival aspect.  Note the two plants with the arching leaves on the left and right sides of the photo above; those plants are Maximillian Sunflower (Helianthus maximilliani), which grew from seed sown in the initial seeding in January.  They flowered their first growing season, a feature not commonly seen in native perennial plants. The large, broad-leaved plant in the foreground is Cup Plant (Silphium perfoliatum), which I grew while working in Nebraska. A fine living memory of a fine state.  Unfortunately, Cup Plant generally flowers in July and August, a period in the Western Cross Timbers of little to no precipitation and drought-like conditions.  I don’t expect Cup Plant to persist here.  Note the abundance of Crabgrass, still.  Better to have that than Bermudagrass.

By the end of August, the Maximillian Sunflowers had grown to 6′ in height and just started to flower.  Other species beginning to flower were Curly-cup Gumweed, Sky Blue Sage, Texas Greeneyes, Big Bluestem and Switchgrass.

October – An Odd Surprise

Tradescantia occidentalis

For whatever reason – a questionable seed source from a Texas-based seed dealer, a dying plant, or just the right environmental conditions at an odd time, Prairie Spiderwort (Tradescantia occidentalis) flowered.  Whether this was a bad omen of things to come or a hopeful, positive sign of a future reconstructed prairie remains to be seen.  After the flower withered away for the year, I checked the fruits left behind; they were all aborted.

A search in the Flora of Texas Database showed no Spiderwort entries flowering later than August 28 (El Paso Co.).  Likewise, a search through Google Scholar turned up no mentions of fall flowering for the species in Texas, nor does Correll and Johnston’s landmark 1970 publication make any mention of such a late flowering time for the species.


The young prairie entered its first season of dormancy not long after the first cold front of fall roared down the plains on a late October day.  I was pleased with the prairie’s first year, but realized that the absence of Little Bluestem – the chief graminoid the prairies of Texas – left a huge hole in the aspects and feeling of completeness to the young prairie.  This year I will make sure Little Bluestem gets established via seed and greenhouse-grown seedlings.

Observations Made and Lessons Learned

Ideally, an easier way to convert Bermudagrass lawn to prairie is to spend a minimum of one full year using chemical and mechanical site preparation. A minimum 20% glyphosate solution is initially needed to severely set back Bermudagrass, and timing is critical to allow the chemical to achieve maximum kill rate.  More than one application may be needed. Mechanical site preparation will be needed, such as light discing or shallow plowing using sweeps on a field cultivator.

The next best step is to plant seed of wildflowers that have heavy hard seeds and are able to establish rapidly, such as:
American Basketflower (Centaurea americana)
Texas Bluebonnet (Lupinus texensis)
Texas Thistle and Wavyleaf Thistle (Cirsium texanum; C. undulatum)
Partridge Pea (Chamaecrista fasciculata)
Common, Plains, and Maximillian Sunflowers (Helianthus spp.)
Black-eyed Susan (Rudbeckia hirta)
Clasping Coneflower (Dracopis amplexicaulis)
Prairie Coneflower (Ratibida columnifera)
Engelmann Daisy (Engelmanniana peristenia)

Some slower to establish perennial species can be added as well:
Narrowleaf Coneflower (Echinacea angustifolia)
Engelmann Daisy (Engelmanniana peristenia)
Rosinweed and Compassplant (Silphium spp.)
Stiff Sunflower (Helianthus rigidus)

Most of those species listed above are available commercially, either as selected native germplasms,  (a list of appropriate commercial seed sources for plantings in NCTX can be seen here),  or as part of a “wild harvest”.  Buy wild harvested seeds with caution! Spotted Knapweed (an invasive European weed) is now in Texas because of the use of wild harvested seed.  The species listed above can also be harvested locally by hand – a rewarding and educational experience for those seeding a small area.

Patches of Bermuda and other exotic grasses may routinely appear in scattered locations across the planting site.  An herbicide that is grass selective can be sprayed over the top of the established forbs with no harm done except to the targeted grasses.  No later than the second season, a light discing followed by overseeding of grass species can be done to complete the initial seeding.

Plan to spend 3-5 years of planning, site preparation, seed collecting (or purchasing), monitoring and spraying, and overseeding.  Even then, the work isn’t finished.  Continued overseeding with less common species should be utilized to increase species richness and give power to the dynamics of secondary succession and the prairie matrix, along with the occasional application of herbicides and the re-introduction of fire.

For those used to and expecting a more tame appearance, like the oft used mass plantings of single species groupings, all young prairies will be “weedy” before they are pretty.  Secondary succession won’t let you have it any other way.

Reminder for Spring Wildflower Season in TX

As spring wildflowers begin their annual show across Texas, it would be helpful for all of us to be on the lookout for Spotted Knapweed (Centaurea stoebe), the latest exotic invasive arrival in the Lone Star State. I wrote last year that Spotted Knapweed has never been found in Texas before and now it is here because of seed contamination susceptibility of the wild harvest model (WHM) used in some grassland restorations. Enjoy this year’s wildflower show – it should be a banner year in Far West Texas. Continue reading

The Vernal Desert

In Texas the beloved and common indicator of springtime’s arrival is the Bluebonnet.  The wildflower is also the hallmark of spring road trips in the state.

There are six species of these members of the genus Lupine in the state and all are considered to be the state flower (state law also grants that any new species discovered will be recognized as an official state flower), thus Texas has the distinction of actually having six state flowers. Continue reading

The Old Stuff

In the days of the old prairie, its members reveled in dormancy and cast their seeds with a starry eye and a mind full of hope toward the next vernal season.  Returning each growing season only from seed, those with a monocarpic life cycle, such as the obligate annuals, have no rooted memory of their hard work; only genetic memory contained in the germ left behind.  The perennials however, are deeply rooted with a semi-permanence of interred vegetative memory and are much longer lived; some individual clonal species may be several decades or even centuries old.  Some perennials may flower many times and produce an abundance of seed throughout their life.  Others may flower and set seed just once, even after living for many decades. Continue reading

The Discovery of Spotted Knapweed in Texas: Why Seed Origin Matters

In the summer of this year, Spotted Knapweed (Centaurea stoebe), for the first time ever, has become the latest exotic member of the flora of Texas.  Its seeds arrived as a contaminant of a wild harvest originating in the Central or Northern Great Plains, which was purchased by a Texas-based seed dealer. This introduction has many negative ecological implications for the state as a whole, and highlights the many risks and flaws that are inherent and eternal to the wild harvest model (WHM).  The initial discovery was made at the Balcones Canyonlands National Wildlife Refuge in Travis County.

Because almost all of the native prairies in Texas have been plowed up and converted to “improved” pastures full of exotic species (species that are noxious weeds, regardless of legal listings), and all remnant (unplowed) prairies have been invaded by exotic plant species, some seed dealers elect to supplement their seed stocks by purchasing bulk “wild harvests” from northern seed growers. This is done in addition to harvesting their own locally grown seed stocks. However, those wild harvests are within the distribution ranges of many northern exotic species like Spotted Knapweed (C. stoebe), Dame’s Rocket (Hesperis matronalis), and Smooth Brome (Bromus inermis), and once exotic seed is in the harvest mix there is no getting it all out. A handful of exotic seed is all it takes to start an invasion.  This is but one example of many why the WHM has no future in TexasThat model carries too much ecological risk, namely in facilitating exotic plant invasions via non-certified seed.  Unfortunately, there are companies who do wild harvests in Texas and sell that seed, which carries equal risk of noxious weed seed contamination, especially the exotic grasses KR Bluestem, Johnsongrass, and Buffelgrass.

Grassland restoration in Texas is largely implemented using the select native germplasm model (SNG), first used by the Soil Conservation Service in 1935, now the NRCS Plant Materials Program. The SNG model is being further developed and used with great success by South Texas Natives and Texas Native Seeds (Central and West Texas), modeled after the highly successful Iowa Ecotype Project (now the Natural Selections Program) at the Tallgrass Prairie Center of the University of Northern Iowa. South Texas Natives (STN), a part of Texas Native Seeds (TNS), is a partnership led by the Caesar Kleberg Wildlife Research Institute (CKWRI) of Texas A&M-Kingsville, working on developing certified ecotypic native seed sources, and advancing restoration research and implementation. Texas Native Seeds operates their research and development in South, West, and Central Texas at this time. Certified native seed is the fundamental aspect of their initiative and model.

With their model, STN has ensured seed for restoration through 25 ecotypically-appropriate plant releases. Presently, certified seed to re-seed about 30,000 acres annually is available to restoration consumers through cooperating commercial seed companies. Aside from the fact that such seeds are a far better choice simply because of certification standards and a very high quality of cleanliness and germination, the wild harvest model will never match that capacity year in and year out. Nature and land fragmentation work against the WHM — drought and lack of quality harvest sites being the major limiting factors — forcing WHM companies to supplement seed stocks with commercial varieties of native seed.

A major economical positive of the SNG model is the lower price point for consumers and a higher profit margin for the producers.  More seed per species per acre can be grown using selected germplasms than can be harvested using the WHM.  Non-selected germplasms (considered maladapted ecotypes until proven otherwise) generally have a much lower germination rate than selected germplasms, which have consistently higher germination rates and higher genetic purity — which can be readily maintained for best ecotypic use — than the wild harvested seeds, whose genetic purity cannot be maintained (or even known), simply as a fundamental flaw of the WHM. Often, wild harvested seeds are of lower genetic diversity than SNG seeds because most wild harvests in Texas come from a few small, locally isolated stands. By contrast, most SNG seeds are comprised of multiple populations (multiple genetics), with the hope of making the seed products adaptable across a larger region and across multiple soil types.

On the issues of the inherent and eternal flaws of the WHM, some might ask, “Why care about which model contains the most or least flaws? What does it matter?” It matters because we have Spotted Knapweed in Texas. Spotted Knapweed has never been in the state before and now it is because of the high risk of seed contamination associated with the WHM. The list below highlights many of the major faults inherent and eternal to the WHM.

The Inherent Faults and Eternal Flaws of the Wild Harvest Model

  • Seed production is highly variable. Some years the production and quality are the mother-lode; other years it’s a bust. Such inconsistent seed production and quality makes wild harvests unpredictable and costly, leading many wild harvest-based companies to supplement their seed stocks with commercial germplasms during years of lean seed production in “wild” stands. Consequently then, lack of supply also leads to use of exotic grasses in mandated seedings.
  • Origin(s) of WHM seeds are often unstated. While some companies may tell you the origins of the harvest or the seed lots, many will not. Beware of those who refuse to give out that critical information. You, as a consumer, have a right to know where the seed originated from!
  • Wild harvested seed is often 2-3 times more expensive than select native germplasm seed. As a straight comparison, based upon pure live seed (which is unstated for wild harvested seed), certified SNG seeds may cost as little as 1/10th the price of wild harvested seed. Your dollar goes further and covers more acreage using the SNG model.
  • Because the wild harvest model is limited to sites that are large enough to make it cost effective for the use of combines and labor, so too are the genetics within each site limited.
  • The issue of exotic species contamination is huge with the wild harvest model. The Spotted Knapweed fiasco illustrates this very clearly: because of the many inherent holes in the WHM, we have Spotted Knapweed in Texas today!
  • The vigor of plants in the populations from which the seeds are harvested is unknown. Plant vigor and performance is critical to restoration success and competitive ability amongst the many exotic species in the state.
  • Areas of adaptation of wild harvested seed is not known. They are untested and unproven in many areas. Local-only purists are often against evaluation of plant adaptations and performances because they believe that scientific aspect somehow doesn’t fit in with their narrowly-focused narrative of “true” prairie restoration.  Evaluations prove potential restoration success on a species by species basis, and certification proves origin. In most SNG seed sources, great care is taken to not alter the original genetic makeup of wild populations.

Following the last point, many of the criticisms placed upon the SNG model by the local-only crowd (inappropriate genetics, potential for genetic swamping, creates larger seeds, current prairie genetics cannot be improved upon, etc.) ignore the fact that severely degraded sites require much more intervention in order to implement the restoration of the plant community that existed prior to the degradation. Simply harvesting seeds with a combine in a “wild” prairie and then scattering them on hyper-degraded sites sets up ecological bankruptcy, wasting time, money, and hard effort.  Restoration today is not as simple as broadcasting unproven seed onto the ground; it requires dedicated research of extensive available literature, replicated evaluations at multiple sites, and further testing in demonstration plantings.  Jones (2013a) expounds upon this when he writes, “While these arguments [that local is best] have considerable merit, they may be less applicable for seriously degraded lands. The best management practices for sustaining mostly pristine lands may differ from those for restoring novel ecosystems. This is particularly noteworthy because such ecosystems are expanding at the expense of pristine landscapes worldwide.”

When it comes to persistence amongst exotic species, SNG seeds have the upper hand in germination-competitive abilities over the unknown track record of wild harvested seed.  As stated by Jones (2013b), “a failure to recognize the importance of augmented genetic variation in restoration plant materials demonstrates a lack of confidence in the reality and utility of natural selection”.  Furthermore, Jones (2013b) states, “one cannot continue to rely solely on local genotypes [simply] because they are local and theoretically best adapted if experience demonstrates otherwise.”

High quality, clean seed is the key to restoration.  No matter how carefully the seedbed preparation and planting is done, using wild, maladapted seed sources will usually lead to planting failure.  Often, those who experience failure when using questionable seed sources do not try again with native species but instead use exotic species, which further perpetuates many of the problems grassland ecologists are dealing with.  Certified native seed of selected native germplasms is the only way to go in helping to maintain native plant communities against exotic species increases and implement successful restoration plantings.

In summary, if the seed you are looking to purchase isn’t certified native seed as regimented under a state agricultural program, if no origin information is given or the seed grower won’t or “can’t” tell you the origin (listing county of origin does not in any way give up a collection source), if it’s not sold on a Pure Live Seed (PLS) basis and only on a bulk pound basis with no germination rate(s) given, then it’s a crapshoot and is anyone’s guess as to what really germinates in a planting.

And thanks to the many holes within the wild harvest model, we now have Spotted Knapweed in Texas. The use of certified seed could have prevented this!

If you have purchased non-certified native seed from a Texas-based seed dealer at any time in the last two years, you are at risk for having Spotted Knapweed in your planting(s).  Many ecologists expect Spotted Knapweed to show up outside of Travis County, especially along roadsides and in private plantings where its population increases are expected to be highest, simply because private landowners are less strict in their monitoring.


Jones, T.A. 2013a. When local isn’t best. Evolutionary Applications. 6:1109-1118.

Jones, T.A. 2013b. Ecologically Appropriate Plant Materials for Restoration Applications. BioScience. 63(3):211-219.

From left to right, top to bottom: Plant view of Spotted Knapweed; flower; flower bud; midstem leaf; and basal leaves.  Photos courtesy of USFWS Balcones Canyonlands.
From left to right, top to bottom: Plant view of Spotted Knapweed; flower; flower bud; midstem leaf; and basal leaves. Photos courtesy of USFWS Balcones Canyonlands.  Do not confuse Spotted Knapweed with the native forb American Basketflower (Plectocephalus americanus [new name] Centaurea americana [old name]).