Yellow Stargrass (Hypoxis hirsuta):

Seventh Year Report on Experiments to Restore a Conservative Plant

April 2013 to June 2019

We wanted answers to these questions:

- Can we restore Hypoxis (“yellow stargrass”) effectively through dormant corms?

- If so, will successfully restored plants reproduce themselves in a competitive, conservative ecosystem? If so, at what rate?

- Will Hypoxis transplants thrive in all open grassland habitats, including those rank with aggressive native species?

- Will Hypoxis thrive in the dappled shade of savanna/woodland oaks? 

- Can briefly trained volunteers do this restoration work adequately?

Spoiler alert: Here are some thumbnail answers, preliminarily suggested by the data.

Yes, briefly trained volunteers, mostly during one morning in 2013, successfully restored at least 148 plants of Hypoxis, which we re-found in 2019. Transplants thrived in fair to higher-quality grassland vegetation (mesic, wet-mesic, and in one case a plant mostly under water as we monitored it) but not in rank vegetation or under trees. Little apparent reproduction by the restored plants was found. 

Small, but possibly mighty.
(Photo by Lisa Musgrave.)

Introduction

Yellow stargrass (Hypoxis hirsuta) is said to be “common throughout the tallgrass region in prairies ranging from dry to moist, as well as in open savannas and woodlands.” (Ladd 1955). Indeed, frequently this plant is so common in high-quality grasslands that its flowers speckle every square foot. Betz lists it with only eight other species as comprising the highest stage of prairie succession or recovery. This species is given a conservativeness (“high quality”) rating of 9 out of 10 for the Chicago region by Swink and Wilhelm. It is a plant in the 7 to 10 range throughout the eastern tallgrass prairie and savanna region. See: Universal FQA. 

In 1985, Hypoxis was absent from most of  our four-decade-old restoration-experiment site, Somme Prairie Grove, a prairie and savanna complex in the Cook County Forest Preserves in Northbrook, Illinois. Hypoxis was found in only a few better-quality patches (totaling an acre or two) of this 85-acre site. By 2013, those areas had expanded slightly, and dense Hypoxis had spread to three small, new areas – those where seed broadcast had introduced many conservative species, thanks to early, difficult seed collection by a few heroes. But, in 2013 the stewards no longer chose to devote much time to such challenging seed collection, for apparently modest benefit. As Hypoxis may be important to the restoration of 80 acres of quality prairie and savanna, we sought more efficient means.  

Here in recovering savanna, yellow stargrass blooms with violet wood sorrel, and common blue violet. Identifiable through their foliage are Solomon’s seal, shooting star, bastard toadflax, nodding onion, Culver’s root, and woodland sunflower. 

Methods

In 2011, we mapped areas where Hypoxis was abundant, scattered, or missing. Results are shown on the map below.

For more detail on this map, see Endnote 1.

We had for some years been growing Hypoxis in a bed devoted to it alone, which by 2013 contained many hundreds of plants. Shortly before transplanting, we dug up more than two hundred corms ranging in size from smallish acorns to small peas.  Many of the corms already had sprouts about ¼ inch long. We would have preferred to transplant them while completely dormant. 

On the 21^stof April 2013, volunteer stewards planted Hypoxis corms in eleven of the fourteen transects monitored for this report. For transects along paths, one corm was planted on each side of the site’s narrow footpaths, at intervals of ten meters (measured by pacing and marked for the planters by colored flags). For two transects away from paths, volunteers were asked to plant one corm at each flag. 

Classic introvert photo of the backs of volunteers planting star grass corms
along a flagged transect on April 21, 2013.

On the map below, planted and monitored transects along paths are shown by red outlines. Transects off trails are shown with lines of red x’s. Transects H, I, J, and K were through poor quality vegetation. Transect F and G were not included in this year’s monitoring, because unburned. See Endnote 2 for more map details. 

Results

How We Recognized Transplants

Counted plants were in areas where no Hypoxis had been seen in 2011. They appeared ten meters apart, often in pairs on opposite sides of the footpaths, as mapped. On most transects, no other Hypoxis plants were seen.  

Success Rate of Transplants

In the case of transects through woods or poor quality (rank) vegetation, we did not find surviving plants and thus did not even precisely locate the transects. It is possible that plants were there and not blooming, or even were blooming but obscured by the rank vegetation. Later we did find one plant along Transect J, the right distance from the path, in the best quality vegetation along that transect. We carefully searched for another plant opposite it – or ten meters in either direction – but found no more. 

On May 23 through May 28, 2019, we found fourteen transects in good quality restoration areas and counted 148 successfully restored plants. This effort’s calculated success rate for those transects is 66%. For math, reasoning, and other details, see Endnote 3.

Reproduction from Restored Plants

On most transects we observed no plants of Hypoxis other than the individual, initial restored ones (about half of them mature plants when restored). Thus, six growing seasons later, we found no apparent reproduction at 120 of those 122 points. For two of the points, where we found a few additional plants nearby, our 2011 map showed one or two plants already present. In some of those cases, where there had been one or two plants, there were now ten or fifteen plants, often in more or less of a line. Perhaps those plants had already been dispersing seeds in 2011 for some years, and those original plants had by now matured.  It is entirely possible that seedlings have arisen from our transplants but are still immature and hard to find. Future monitoring will be needed to determine if and when the 120 documented-as-restored plants successfully reproduce.

Stargrass "as thick as grass" - where apparently seeded decades ago - along with seeded compass plant, purple prairie clover, rattlesnake master, wood betony, and cream false indigo. 

Discussion

A natural prairie or savanna with full biodiversity depends on species of plants that bloom in spring, summer, and fall. Summer and fall blooming species are more readily restored than spring blooming species. Yet, thanks to diligent seed collecting by many, some formerly-absent, characteristic plants of high-quality open grassland spring flora have increased dramatically in Somme Prairie Grove. These include such conservative species as: Baptisia leucophaea, Dodecatheon meadia, Panicum leibergii, Pedicularis canadensis, and Phlox pilosa. Other species are proliferating in limited areas but (perhaps because seed is more difficult to collect) are absent from large areas; these include Comandra umbellata, Hypoxis hirsuta, Lithospermum canescens, Scutellaria parvula, and Viola pedatafida). One spring species, Heuchera richardsonnii, has established new populations only marginally, despite much seed broadcast. It is our impression that conservative species facilitate the recovery of others; many quality species do not thrive among invasive or other aggressive species. 

Might restored Hypoxis and similar conservative species in time promote the expansion of high-quality and reduce rank vegetation? As Hypoxis seems not to establish in poorest quality areas, might other species need to prepare the community in some way before Hypoxis can succeed, as described by Betz (see Endnote 4)? Future monitoring and experiments are needed. This experiment demonstrated, so far, that Hypoxis can be established by transplanted corms in recovering prairie and savanna that began as old-field vegetation, enriched by inter-seeding of prairie and savanna species. 

More of this kind of work could aid biodiversity conservation by expanding remnants of sub-viable size. See Endnote 4. 

ENDNOTES

Endnote 1

The map below shows the results of Hypoxis monitoring on May 20, 2011. This level of monitoring is not as accurate or complete as an excellent grad student might do for a thesis. But it’s what we had time for, and we hereby argue that it is good enough to learn from. (We recommend that serious, long-term practitioners find time and resources for at least this level of monitoring.) This and similar studies can be improved or corrected by more detailed work, but working hypotheses can tentatively, reasonably be established in this way. 

On May 20, 2011, I had walked trails and mapped Hypoxis as to whether it was dense, scattered, or absent. Green outlines indicate areas of dense or scattered Hypoxis. Green stars indicate individual, lone plants. Lines of green dots away from trails indicate off-trail areas that I censused. Red outlines show areas where I found no Hypoxis. Young or other non-blooming plants would not have been found in a survey of this kind.

A re-check on June 1 found some plants that had been missed in May. On what would be Transect M, a group of five plants was missed, and on Transect D, three plants were recorded where only one had been found twelve days before. Earlier searches have the benefit of less obscuring tall vegetation; later searches have the benefit of more individual Hypoxis plants in flower. On the June 1 map, Hypoxis is in green, and Oxalis violacea is in red. The lack of plants is indicated by “0.” If a few plants were seen in an area, the number of plants is given. The reference to “burned” on this map indicates that this survey covered the southern half of the site, which had been burned prior to the start of the 2011 growing season. 

Endnote 2

More details connected to "Planted Yellow Star Grass" map from 2013 (three maps above, unfortunately): For this report, we did not include Transect F as it turned out to have been put through difficult, rank vegetation in part, and in part through areas that already had dense, unrestored Hypoxis. It represents a hasty, poor decision, late in the day. 

Transect G was not monitored for this report, because it ran through an area that had not been burned this year (the west half of the site was burned) and had no visible Hypoxis on May 23. 

Three transects were noticed that had not been on the April 2013 map. These are shown below as Transects X, Y, and Z. 

We have as yet found no maps or notes for these three transects. But there they were - Hypoxis every ten meters. We decided to include them in this report as the long Z transect and part of the X transect were burned this year. They were probably planted in 2012 or 2014 – years when the north half was burned.  They will now not be burned for two years. (We are burning half the site each year, following the pattern: west half, south half, east half, north half.) The above map also shows the poor way that data was recorded when we ran short of paper.  

Endnote 3

Variable rates of success can be expected on the basis of weather, quality of material introduced, skill and dedication of the crew, disease, disturbance by animals, and many other factors. In our experience, digging up by voles is a major threat, for which reason we now install transplants predominantly in recently burned areas. (Voles flee from and are heavily predated in burned areas for at least the first few weeks of the growing season, after which our transplanting disturbance is much less obvious and attractive to them, thankfully.)

Because 104 points represented two planted individuals (one on each side of path) but 18 points on non-path transects represented one planted individual each, with 100% success we would have expected 224 plants (226 x 2 + 18 - 2). Since we found 148 plants, our observed success rate was 66% (148/224). 

The observed success rate is likely lower than the real rate. These plants are difficult to see among other vegetation when not in flower. Although we did find and count some non-flowering plants, these were all in bud and along a path, exactly opposite an observed flowering plant. On the non-path transects, we initially found no points or plants at all. (Landmarks referenced in notes were not as clear as they could have been.) After a few tries, we found the right landmarks and, to our surprise, started seeing a Hypoxis most every 11 or 12 paces as we headed toward a now-dead elm, or, fortunately, in most cases, a living oak. (Packard’s pace is 0.86 meters. Pacing in a high-quality grassland is made less accurate by hummocks and the disinclination to step on rare plants. But it turned out to be accurate enough for this purpose.)  

A large proportion of the plants found seemed exactly across from each other on opposite sides of the paths – despite the fact that people needed to deal with stumps, crayfish holes, thorny roses, and other challenges. Plants were not all at equal distances from the footpaths, but most pairs were at equal distance. This phenomenon seems to reflect that fact that the stewards knelt in the path at the flag and planted the corms perpendicularly at the convenient distance for a person of that height. In other words, the corms planted by a taller person ended up farther from the path than those by a shorter person. No problem. 

Some wrinkles required judgment. At one point, a large dead stump was opposite the found Hypoxis. No Hypoxis could have been planted there. We subtract one from the expected number of plants, as we did for another point that had long been unplantable, as dense brush (thus, the “-2” in the calculation above). All other questionable points we treated as if both sides of the path had been planted. 

In the case of the non-path Transect B, my memory and notes suggested that this was a “single corm” transect like E and F. We counted 12 plants at 8 points, but oddly, one point had two plants, and one had three, straddling the transect in a straight line. Was this the end of the day, and some tired, creative person was trying to use up the plants in a reasonable way? That’s what we guessed. We included these three “extra” plants in the calculation. Perhaps we should have subtracted them (success rate then 64%), or also added the three to the “expected” side of the equation (success rate then 65.6%). But, really, these numbers are iffy in the extreme to start with. What about plants eaten by bugs, or stepped on by a deer, or just not blooming? The usefulness of this data does not depend on that level of precision.   

Endnote 4

Dr. Robert Betz, the principal scientist and sparkplug behind the ecosystem restoration movement in the Chicago region, taught and wrote that an effective way to establish prairie on recently plowed cornfields was to initially plant mostly “first wave” (easily established) species and later focus on “second, third, and fourth wave” species. Betz lists nine species as comprising that ultimate fourth wave, one of which is Hypoxis hirsuta.

Yellow stargrass growing with a fellow high conservative, the prairie lady-slipper.

At Somme our work has been, not on recently plowed cornfields, but on former pasture or former cornfield that through years as forest preserve had succeeded to “old-field” vegetation. In the case of three of Betz’s “final stage” species, for which we were able to acquire seed and run experiments, we found that three would establish readily from seed in old-field turf (Platanthera leucophaea and Gentiana puberulenta) or in high-quality prairie (Cypripedium candidum).

As his “third wave” – Betz listed thirteen species, most of which we have found to establish readily in old-field vegetation, when plentiful seed is broadcast. More such work could increase the viability of prairie and savanna remnants generally – by expanding small populations of rare plants and the many remnant-dependent animal species that depend on them.

Names of Plant Species in this Post 

(C = Coefficient of Conservatism)

Scientific Name	C	Common Name
Allium cernuum	7	nodding wild onion
Baptisia leucophaea	10	cream wild indigo
Comandra umbellata	7	bastard toadflax
Cypripedium candidum	10	white ladyslipper
Dodecatheon meadia	6	shooting star
Eryngium yuccifolium	9	rattlesnake master
Gentiana puberulenta	10	prairie gentian
Platanthera leucophaea	10	eastern prairie fringed orchid
Helianthus hirsutus	5	hispid sunflower
Helianthus strumosus	5	pale-leaved sunflower
Heuchera richardsonii	8	prairie alum root
Hypoxis hirsuta	9	yellow star grass
Lithospermum canescens	8	hoary puccoon
Oxalis violacea	9	violet wood sorrel
Panicum leibergii	10	prairie panic grass
Pedicularis canadensis	9	wood betony
Dalea purpurea	9	purple prairie clover
Phlox pilosa fulgida	7	prairie phlox
Polygonatum canaliculatum	3	smooth Solomon’s seal
Scutellaria parvula leonardii	7	small skullcap
Silphium laciniatum	5	compass plant
Veronicastrum virginicum	7	Culver’s root
Viola pedatifida	9	prairie violet
Viola sororia	3	common blue violet

References

Betz, Robert F. 2001. The Prairie of the Illinois Country. 

Ladd, Doug. 1995, Tallgrass Prairie Wildflowers.

Swink, Floyd and Gerould Wilhelm. 1994. Plants of the Chicago Region. 

Wilhelm, Gerould and Laura Rericha. 2017. Flora of the Chicago Region. 

Companion Post

A common-name based, less technical, more fun, but not all that much fun, and trying to be more insightful version of this post is at: https://vestalgrove.blogspot.com/2019/06/yellow-stargrass-humble-and-dominant.html

Acknowledgements

 

Hundreds of people deserve thanks for their contributions as the volunteer stewards and Cook County Forest Preserve staff who have nurtured Somme Prairie Grove over the decades. David Painter deserves admiration for his impressive harvests of seed of difficult-to-gather conservative species. Eriko Kojima helped find and monitor the transects in 2019 and prepared the table of scientific and common names. Kathleen Garness helped correct many typos and whatever you call those annoying changes that spell-check sometimes makes.