“Five Reasons it’s Okay to Love Oxalis – and Stop Poisoning It”

The San Francisco Forest Alliance has published an excellent article about the pointless and harmful attempt to eradicate oxalis in San Francisco’s so-called “natural areas.”  We are grateful to SFFA for their outstanding research and permission to republish the article. 

The oxalis season is over, and the perky yellow flowers have vanished for another year. These Bermuda buttercups will be back next year to herald the spring, bringing joy to those who love them, irritation to those who hate them, and Tier I herbicides targeted at them in San Francisco’s so-called “Natural” Areas.

oxalis 1


These flowers are so visible in spring that Bay Nature magazine did an article about them in March 2015: A Natural History of the Little Yellow Flower that’s Everywhere Right Now. It quoted Jake Sigg, the retired SF Recreation and Parks gardener who is considered the doyen of San Francisco’s native plant movement. He hates oxalis pes caprae, which he considers extremely invasive. The article quotes him as saying that, without intervention, “in X many years Twin Peaks would just be one solid mass of yellow, and there wouldn’t be any other plants there…” The article suggested that an oxalis-dominated  landscape “drives away coyotes, hawks and owls that feed on grassland foragers, and the situation is especially dire for endangered Mission blue butterflies, which depend heavily on native wildflowers.” Most of those ‘facts’ about oxalis are mistaken as we’ll explain below.

Mr Sigg’s theories align with those of the Natural Areas Program (NAP) of the San Francisco Recreation and Parks Department (SFRPD), which uses the herbicide, Garlon (triclopyr) to battle oxalis despite its dubious efficacy for the purpose.  San Francisco’s Department of the Environment classifies Garlon 4 Ultra as Tier I: Most Hazardous. It’s listed as HIGH PRIORITY TO FIND AN ALTERNATIVE (their caps). Since oxalis is the main reason NAP uses Garlonthe alternative we propose is – don’t use Garlon or anything else on oxalis.

An article on SaveSutro.com, based on a detailed study by the Marin Municipal Water Department, describes some of the issues with Garlon:

  • Garlon “causes severe birth defects in rats at relatively low levels of exposure.” Baby rats were born with brains outside their skulls, or no eyelids. Exposed adult female rats also had more failed pregnancies.
  • Rat and dog studies showed damage to the kidneys, the liver, and the blood.
  • About 1-2% of Garlon falling on human skin is absorbed within a day. For rodents, it is absorbed twelve times as fast. It’s unclear what happens to predators such as hawks that eat the affected rodents.
  • Dogs  may be particularly vulnerable; their kidneys may not be able to handle Garlon as well as rats or humans.  Dow Chemical objected when the Environmental Protection Agency noted decreased red-dye excretion as an adverse effect, so now it’s just listed as an “effect.”
  • It very probably alters soil biology. “Garlon 4 can inhibit growth in the mycorrhizal fungi…” (soil funguses that help plant nutrition)
  • It’s particularly dangerous to aquatic creatures: fish (particularly salmon); invertebrates; and aquatic plants.
  • Garlon can persist in dead vegetation for up to two years.
Natural Areas Program uses Garlon on oxalis
Natural Areas Program uses Garlon on oxalis

First, a little about the actual natural history of oxalis. This plant doesn’t set seed in California, and spreads entirely by sending out roots and forming little bulbils (like tiny potatoes) underground. It’s usually found where the soil has been disturbed by activities such as road-building, gardening, or trail-building. In some cases, the disturbance comes from landslides or something similar. It can’t stand frost. If we do nothing, it would tend to die down rather than spreading uncontrollably.

In disturbed landscapes, it can spread fast. For this reason it can be a nuisance in gardens. People don’t want to leave their gardens alone for years to let nature take its course with the oxalis, and not every garden design includes brilliant yellow as the dominant color for a few weeks. The only way to eradicate it in the short term is to dig it out carefully every time you see it, and make sure you get most of the bulbils. Or use strong herbicides, which may not work.

In a natural landscape, though, it’s a different story and here’s why.


Honeybee on oxalis flower
Honeybee on oxalis flower

Oxalis is actually an excellent plant for bees and butterflies.  When blooming, it provides “copious nectar.” In fact, it generously gives away its nectar. Since it doesn’t set seed, it doesn’t benefit from pollinators – but it’s a food source for honey bees, bumblebees and butterflies. (You can read a rather technical description of the plant HERE in a 2-page PDF note from UCLA’s Barry A. Prigge and Arthur C. Gibson.)

Bumblebee on oxalis flower
Bumblebee on oxalis flower

In fact, a recent 2014 study shows that plant communities with exotic plants had more plant species as well as more pollinators, that pollinators didn’t prefer native plants, and that even some specialist pollinators depended on introduced plant species.

It’s true the Mission Blue butterfly needs (native) lupine as its nursery plant. (It doesn’t depend on any other native wildflowers – only three varieties of lupine.  Incidentally, one of the key nectar sources for the Mission Blue butterfly is an invasive non-native Italian thistle: Carduus pycnocephalus).

Butterfly on oxalis flower
Butterfly on oxalis flower

Lupine has been planted on Twin Peaks as NAP attempts to reintroduce the Mission Blue butterfly there. But lupine is also a plant of disturbed areas, which means that NAP must maintain it or it will die out as the area stabilizes. They have to keep planting it, weeding, and trimming the grass around the lupine patches to make it attractive to the butterfly. An SFRPD report on the reintroduction project said “unmanaged habitat deteriorates quickly.” Presumably, they don’t use Garlon near the lupine patches, since it would likely kill that too. Despite what is implied in the Bay Nature article, it’s not oxalis that’s the issue. The real problem is another native plant, the coyote bush which takes over grasslands in a natural succession.


Oxalis bulbils are a food source for wildlife. Gophers and other rodents eat them. In fact, the Bay Nature article says, “Their spread is abetted by pocket gophers and scrub jays, which have been spotted carrying the bulbs and caching them in the ground—effectively planting them in new areas.”

Since gophers are a foundation species in the food web, being dinner for predators from hawks to coyotes to great blue herons, these plants actually provide habitat benefits whether or not they’re flowering, because the bulbils are there all year.

oxalis 6

Where there are gophers, the predators follow. Like the coyotes in these pictures, which clearly haven’t been driven away by a landscape dominated by oxalis.

coyote pouncing in oxalis field. Copyright Janet Kessler
coyote pouncing in oxalis field. Copyright Janet Kessler
coyote in oxalis field. Copyright Janet Kessler
coyote in oxalis field. Copyright Janet Kessler


The article says that oxalis leaves “bare ground during the six months of the year oxalis doesn’t flower.” That’s not true either.

oxalis 9

oxalis 10The spectacular yellow bloom of the oxalis – valuable because the mass of color attracts honey bees and bumblebees – gives the impression that it’s the only plant there.  But though it visually takes over the landscape when it’s in bloom, it naturally grows interspersed with grasses and other plants. Like in the picture above.

In fact, oxalis tends to enrich the soil with phosphorus, which is good for grass.

So when it finishes blooming, as it has by now – you don’t get bare ground. The picture below shows the same area as the first picture in this article – but it’s after the oxalis bloom is over. It’s grassland.

oxalis 11


One argument – related  to the ‘bare ground’ argument – is that oxalis takes over grasslands and destroys them, particularly the native grasses. However, grasslands in most of California including San Francisco are dominated by non-native grasses. The change occurred over 100 years ago, when these grasses were planted for pasture. So the grassland that NAP is defending with herbicides is primarily non-native anyway.

oxalis 12

But anyway, what’s the evidence that oxalis is actually damaging native plants?

It’s true some European studies do suggest that an increase in oxalis is associated with a decrease in native plants diversity –though whether it’s a cause is unclear. It may just be benefiting from human activities that disrupt the landscape. Another study put oxalis head-to-head with a native annual grass, lolium rigidum. The native grass tended to dominate. Their conclusion: “Oxalis is a poor competitor. This is consistent with the preferential distribution of Oxalis in disturbed areas such as ruderal habitats, and might explain its low influence on the cover of native species in invaded sites.

The California Invasive Plant Council rates its invasiveness as “moderate,” considering it as somewhat invasive in sand dunes and less so in coastal bluff areas.

In San Francisco, every place where oxalis grows is already a disturbed environment, a mix of non-native grasses and plants with native plants (some of which have been artificially planted).  Here,  oxalis appears to grow happily with other plants – including, for instance, the native California poppy in the picture above.


Children love oxalis, both for its pretty flower and for the sour taste of its edible stems.

oxalis 13

oxalis 14Even small children love gathering posies of Bermuda buttercups (though picking flowers is technically prohibited in Natural Areas). The flowers are surprisingly hardy for wildflowers, and in a glass of water last quite well as cut-flowers.

The plant is edible, and its tart leaves make a nice addition to salad. People enjoy snacking on its sour stems. Besides Bermuda buttercup, it’s also called ‘sourgrass’ and ‘soursob.’ It does contain oxalic acid (as does spinach, for instance), and so you probably wouldn’t want to make a meal of it. Though in South Africa it’s made into soup.

Adding Garlon to it is probably a bad thing.

Photo credit: Badjonni (Creative Commons – Flickr)
Photo credit: Badjonni (Creative Commons – Flickr)


From our current evidence, there’s no sign that oxalis has a negative impact on wildlife, and plenty of evidence it’s already part of the ecological food web of our city.  The evidence also suggests it’s not having a negative effect on other plants in San Francisco either. Lots of people find this flower attractive; one writer described it as the city smiling with Bermuda buttercups.

In any case, even Doug Johnson of the California Invasive Plant Council doesn’t think it’s worth attacking at a landscape level: the payoff isn’t worth the expense. Removing it from the hundreds of acres in Natural Areas isn’t as simple as eradicating it from a small yard where it’s clashing with the garden design. It requires a lot of work, a lot of powerful herbicides, a multi-year effort – and for what?

The justification for using strong pesticides like Garlon to control it is weak. We call on NAP to stop using Tier I and Tier II herbicides altogether.


Concern about herbicide use: Legitimate or “chemophobia”?

Recently there was a brief dialogue about herbicide use in San Francisco’s so-called “natural areas” in Jake Sigg’s Nature News that was of some interest to those who consider such use a contradiction in a public park designated as a “natural area”:

  •  Jake Sigg:  “Spurious, damaging information being circulated regarding herbicide use in our open spaces:  Mischievous people…are circulating false information…whipping up fears that have no foundation”  (Nature News, February 14, 2011)
  • Reader:  “…Garlon is legally classified as a hazardous chemical.  I am therefore writing to supply you with information from reputable sources.  I ask that in your future communications on this subject, you accurately describe the facts that are known about this chemical.” (Nature News, February 18, 2011)
  • Jake Sigg:  “The chemophobia rampant in this country is primarily based on emotion and anxiety, and does us a great disservice.” (Nature News, February 18, 2011)
  • Jake Sigg:  “The anti-herbicide crazies quickly seize on articles like this NYT one as proof of their contentions…” (Nature News, March 30, 2011)

This dialogue and the positive feedback that Mr. Sigg reported from his readers in support of herbicide use, suggest that herbicides are an important tool for the native plant movement.  They are anxious not to lose this tool in their crusade to eradicate non-native plants and trees.  After researching how much herbicide is being used by the Natural Areas Program, we can understand why they angrily defend its use. 

Herbicide use by San Francisco’s Natural Areas Program

The Natural Areas Program (NAP) reports having used herbicides 69 times in 2010:  36 applications of Garlon, 31 applications of Roundup, and 2 applications of Milestone.  Putting those numbers into perspective, other areas (those not designated as “natural areas”) in the Recreation and Park Department sprayed Garlon, the most hazardous chemical, only a few times in 2010.

 About 20% of these herbicide applications were done by a contractor who was paid $9,000 per application.  The employees of this contractor are therefore equally committed to this source of revenue, contributing to the economic interest that is a motivating factor in the native plant movement.

Not all of the “targets” of these sprayings are identified, but those that are include:  oxalis, blackberry, ivy, fennel, cotoneaster, hemlock, pampas grass, broom, erharta grass, mustard, and thistle.  Blackberry is an important source of food for wildlife in the city. We hope that children in the park do not graze on the blackberries.  Garlon was also sprayed on Scabiosa, which has not been identified as an “invasive plant” by the California Invasive Plant Council.

Glen Canyon, with a creek running through it and a year-around day care center adjacent to that creek, was sprayed 12 times.  Twin Peaks, the watershed to that creek, was sprayed 16 times.  Lake Merced was sprayed 3 times, despite the fact that it has been officially designated as red-legged frog habitat.

What is known about these chemicals?

The City’s policy regarding “Integrated Pest Management” classifies the chemicals used on city properties in terms of the risks associated with their use.  Here is how the City’s policy classifies these chemicals: 

  • Garlon:  Tier I, Most Hazardous.  Use Limitations:  “Use only for targeted treatments of high profile or highly invasive exotics via dabbing or injections.  May use for targeted spraying only when dabbing or injection are not feasible and only with use of a respirator.  HIGH PRIORITY TO FIND ALTERNATIVE.” 
  • Roundup:  Tier II, More Hazardous.  Use Limitations:  “Spot application of areas inaccessible or too dangerous for hand methods, right of ways, utility access, or fire prevention.  Use for cracks in hardscape, decomposed granite and edging only as last resort.  OK for renovations but must put in place weed prevention measures.  Note prohibition on use within buffer zone 60 feet around water bodies in red-legged frog habitat.”
  • Milestone:  Tier I, Most Hazardous

    Spraying Garlon on Twin Peaks without use of the respirator required by City policy, February 2011

Federal law also requires that chemicals be evaluated by the Environmental Protection Agency (EPA) before being commercially available to consumers.  The EPA conducts a number of tests of toxicity, reports the results of those tests on a mandated Material Safety Data Sheet (MSDS), and classifies the chemical with respect to its relative toxicity.  Here are a few highlights from the MSDS for these chemicals:

  • Garlon 4 Ultra is defined as a “Hazardous Chemical” according to the OSHA Hazard Communication Standard.  “Material is highly toxic to aquatic species” and “slightly toxic to birds.”
  •  Roundup Pro is defined as a “Hazardous Chemical” according to the OSHA Hazard Communication Standard. Toxicological effects in rats:  “decrease in body weight gain; histopathologic effects.”  “Moderately toxic” to aquatic life.

The Marin Municipal Water District (MMWD) quit using all herbicides in 2005 in response to the public’s protests.  They have been engaged in a process of evaluating herbicides for possible use in the future.  In 2008, MMWD contracted for a risk assessment of 5 herbicides they were considering for possible use.  That risk assessment determined that Garlon 4 Ultra is the most hazardous of the 5 chemicals that were evaluated.    MMWD is not considering the use of Garlon in the future. 

Does NAP’s herbicide use conform to the City’s Integrated Pest Management Law?

 San Francisco’s Integrated Pest Management Ordinance makes the following commitments regarding pesticide use on the city’s properties: 

  • 300a “…the policy of the City..to eliminate or reduce pesticide applications on City property to the maximum extent possible.”
  • “The City…shall assume pesticides are potentially hazardous to human and environmental health.”
  • “The City shall give preference to reasonably available nonpesticide alternatives when considering the use of pesticides.”
  • “Consider the use of chemicals only as a last resort.”
  • “This Chapter applies the Precautionary Principle to the selection of reduced risk pesticides and other pest management techniques.”

The City’s ordinance that obligates the City to follow the Precautionary Principle makes this commitment: 

“A central element of the precautionary approach is the careful assessment of available alternatives using the best available science. An alternatives assessment examines a broad range of options in order to present the public with the consequences of each approach. The process takes short-term versus long-term effects or costs into consideration, and evaluates and compares the adverse or potentially adverse effects of each option, giving preference to those options with fewer potential hazards. This process allows fundamental questions to be asked: ‘Is this potentially hazardous activity necessary?’ ‘What less hazardous options are available?’ and ‘How little damage is possible?’”

We do not believe that herbicide use by the Natural Areas Programs meets the standards of either the City’s ordinance about pesticide use or its commitment to the Precautionary PrincipalThese laws are theoretically rigorous, but the enforcement of those laws is not.  The Natural Areas Program is using an herbicide (Garlon) categorized by City policy as the “Most Hazardous” most of the time.  They are using that chemical in sensitive areas in which water can be contaminated and in which children can be exposed.  Their use of that hazardous chemical has increased over time and they have been using that chemical for at least 5 years, perhaps longer (the 2006 management plan for the Natural Areas Program reports the use of this chemical).  If Garlon has not been capable of eradicating in 5 years, the non-native plants that are the target of the Natural Areas Program, it is not likely to do so in the foreseeable future.  

Legitimate concern or “chemophobia”?

Let the reader be the judge.  Given what we know about these chemicals, the frequency of their use, the length of that use, and the locations of that use:

  • Do you think there is reason to be concerned about the herbicides that are being sprayed on our public parks? 
  • Do you think that places that have been designated as “natural areas” should be sprayed with herbicides which are legally and officially designated as hazardous chemicals? 

Destruction of eucalyptus threatens bees

The Pt Reyes Light received a Letter to the Editor in response to its series about the destruction of eucalyptus trees.  The author of the letter explains that eucalypts are one of the few sources of nectar during the winter, that the nectar is vital to the survival of bees over the winter, and that the bees are essential to California agriculture.  The letter was published in the Light on January 6th and is reprinted here with permission:

Think before you cut

Dear Editor,

The recent articles in the Light regarding the Park’s and other’s plans to eradicate eucalyptus from California fail to take into consideration one critical aspect of the need for eucalyptus in the continuation of agriculture in the state.

The common honeybee was introduced to California in the mid-19th century, around the same time as Blue Gum Eucalyptus. Each spring and summer, honeybees gather huge amounts of nectar from flowers and store it in the form of honey so they will have enough food to make it through the winter, when the weather is too cold and rainy and flowers are too few to provide food for the bees. 

In autumn, each hive greatly reduces its number of bees in order to survive the winter on the honey they stored. This is done by the queen laying fewer eggs and thus not replacing the bees that naturally die. Hives of 40,000 to 50,000 bees in summer drop to 10,000 bees in winter.

During December and early January, bees hover in a tight cluster, keeping each other warm and living off the stored honey.  In early January the Queen again lays eggs in ever-increasing numbers each day; larvae and then newly-hatched bees must be fed huge amounts of honey to support rapid growth. The demand for honey increases exponentially and if honey stores are not enough, the hive can starve to death just before warmer, drier weather and its tons of flowering plants arrives. 

But in California we have periods of sunny, warm days, in January and especially February. These allow bees to forage for nectar to supplement depleted stores in their hives and insure their continuation.  But what is blooming in January and February, when bees are in desperate need of nectar plants? Acacia, almond, ceonothus, manzanita, mustard, rosemary and some fruit trees bloom for short periods of time, but their small number and smaller sizes do not always guarantee enough blossoms. And any hard rain or wind can destroy whatever blossoms there are. 

Eucalyptus, on average 100-feet high and 30 to 50-feet wide, has tens of thousands of nectar-filled blossoms per tree.  It blooms throughout California from late January through mid-May, ensuring an abundant supply of nectar for hives at the time of their most critical need.

Prior to the arrival of the honeybee in California, the state population was 1 million people and agriculture consisted of wheat, barley, cattle and sheep, all of which could easily survive without honeybees.  Today, with California growing much of the fruits, nuts and vegetables for the U.S., the honeybee is an intricate part of the continuation of agriculture. With the current problem of Colony Collapse Disorder, the fate of the honeybee is already precarious. Cut down all these Eucalyptus trees and the fate of thousands of hives of bees, and thus the continued pollination of our food crops, may be in serious jeopardy.  Think before you cut them down.

Cathleen Dorinson, Pt Reyes Station

Eucalyptus and Bee, painting by Brian Stewart
Research on Colony Collapse Disorder has identified reduced supplies of nectar as one of many factors in the failure of about 30% of commercial hives per year in the past few years.  Bees, already weakened by chronic exposure to pesticides and reduced food supplies, are unable to recover from the fungi, viruses, and parasites that are rampant in the “global diaspora of organisms.”

eucalyptus honey

Because of the role of pesticides in the death of bees, the eradication of eucalyptus exposes bees to  double jeopardy:  the loss of a major food source during the winter and exposure to the pesticides that are used to kill the roots of the eucalyptus trees.

Garlon with the active ingredient triclopyr, is the pesticide used by most managers of public lands to kill the roots of the eucalyptus after the trees are cut down.  Eucalyptus is a vigorous resprouter.   Unless the stump is poisoned immediately with a toxic pesticide, it will return ten-fold after it is cut down, or in the unlikely event that it burns down, or after a freeze deep and long enough to cause the tree to die back.

Garlon is known to be toxic to bees.  The Marin Municipal Water District quit using all pesticides on its properties in 2005 in response to public protests.  It hired a consultant to evaluate 5 pesticides for potential use in the future.  The risk assessment published in 2008  stated that Garlon was the most toxic of the 5 pesticides studied and that it was the most toxic to bees. The Marin Municipal Water District is presently seeking approval to begin using Roundup again.  It does not propose to use Garlon.

The so-called Natural Areas Program in San Francisco, which is responsible for the care of approximately 1,000 acres of park land ironically called “natural areas” uses Garlon heavily.  About 75% of its pesticide applications (by volume and frequency) are of Garlon.  Could this be a factor in the collapse of several beehives recently reported in San Francisco?

The East Bay Regional Park District used 34 gallons of Garlon in 2008.  How many more gallons of Garlon will be used by these managers of public lands when they cut down the hundreds of thousands of eucalyptus trees which they have proposed to destroy in their official plans?

Once again, we can’t make sense of the destructive actions of those who are damaging nature in the name of “restoring” nature.  In our view, it is a fundamental contradiction.