Are Daylilies Invasive?

This summer we've been mapping Orange Day Lilies (Hemerocallis fulva) along eastern Ontario roads to see what kinds of places they're growing in. This was inspired by the good season this species has had, the thought that we had neglected them in comparison to other alien roadside plants, and comparing them to Lupines (Lupinus cf polyphyllus) in New Brunswick, which are a wildly popular invasive species which spreads along roadsides from plantings at homesites. 

With the spraying that municipalities are doing against 'poison parsnip,'  killing off all the broad-leaved Dicot herbs, there's the real possibility that Day Lilies will be favoured and become even more widespread along roadsides. We saw the first bloom on 19 June, and were are still a few coming out on 10 August.

Wikipedia says: "Triploid... Hemerocallis fulva var. fulva... native to Asia from the Caucasus east through the Himalaya to China, Japan, and Korea... has escaped from cultivation across much of the United States and parts of Canada and has become a weedy or invasive species. It... spreads more or less rapidly by vegetative increase into woods and fields and along roadsides and ditches... dense stands that exclude native vegetation, and is often so common that it is mistaken for a native species."

Triploid plants characteristically don't set seed, because the three sets of chromosomes get scrambled in the the process of meiosis. Daylilies' one-day flowers wither from precisely regulated cell death, and the unfertilized stems shrivel in their turn. But the Yellow Daylily, H. flava, which is mostly seen in cultivation, is diploid. It sets seeds readily and a myriad of varieties have been produced by interspecies crossing and selection. 

Like Cattails, Daylilies are one of those plants which is edible in every part that's not too fibrous to chew, and while the caution exists that they may loosen the bowels, the tubers, spring shoots, young flower stalks, flowers, and spent flowers are edible, perhaps providing a motive for planting colonies.

This July and August we ran four surveys, at ordinary highway speed or slightly less, waypointing as many of the stands we passed as possible. As is usual with these surveys, some stands may be missed while the observer's head is down writing his notes. The first survey was on 21 July from SSW of Forthton to Bishops Mills, along Temperance Lake, New Dublin, Branch, Kyle, Bolton, and Buker roads. The second was on foot in the SW sector of Bishops Mills on 22 July. The third was with Naomi Langlois-Anderson from Monkland to Jessups Falls along Hwy 43, County Roads 20, 9, 24, and Old Hwy 17. The fourth was along Hwy 17 from Meath to Rolphton on 2 August. The narrative of these observations is appended to this post, so the records will be available to future workers.

Along the 21 July transect, there were 32 sightings, sites averaging 1021 m apart, ranging from 67 m - 4250 m, and a standard deviation of 1117 m; clearly not a normal distribution - Poisson or exponential or something. If eight gaps of more than 1 km were excluded, the distance between sightings averaged 504 m. The habitats were 8 paired stands at driveway, 3 single stands near driveways, 2 rows of plants along roads at Conifer hedges in front of homesites, 11 elsewhere in yard or garden, 4 along a wild edge of a yard, and 7 away from homesites (5 on grassy/herbaceous roadside along woods or brushy oldfield, 2 along grassy fields or pastures, 1 along a farm lane, and 150 m of stands along a cornfield; 31% wild sites).

Along the 24 July transect, on the way out the distances between 18 stands averaged 2532 m (183 - 8570m, standard deviation 2552 m). Coming back some of the sites may have filled gaps, or may have been duplicates. The habitats in this more agricultural landscape were 10 sites in yard or garden, 1 along a wild edge of a yard, and 5 away from homesites (550m of plants along grassy roadside through young woods, 2 along grassy fields or pastures, and 2 stands along tilled fields; 38% wild sites, though this would have been lower if all the stands in Dominionville had been individually counted).

On the 2 August survey of Hwy 17 only 7 sites were seen, separated by an average of 11.4 km (166 m - 51.5 km, standard deviation 18.0 km). The habitats were one on either side of a driveway, a single stand along driveway, two elsewhere in yard or garden, two stands on either side of a motel sign, and wild stands along a marsh and along grassy fields or pastures; 29% wild sites.

Bishops Mills is as rich in Daylilies as anywhere, and we're variously familiar with the stands, though the dates of establishment of stands we've intentionally planted on our land for food either were not recorded, or have not yet made it into the database. These are a row planted, sometime in the 1980s or early 1990s, in the sandy 'Upper Garden' site in oldfield near an Aspen plantation, which is now a 3 x 12 metre patch, and a stand in a grassy patch of shallow-soil oldfield being overgrown by Grape vines and shrubs, that was planted sometime in the early 2000s, and is now a 6 x 5 metre stand. A few plants showed up in 2018 among Garlic Mustard, Nettles, & Motherwort, from somewhere into a now-unused Goat yard, and an old row along the driveway of our neighbouring cottage,'Pipers House' is spreading a few metres into an adjacent irregularly-tilled garden area.

There's a 4 metre stand around a telephone pole across the street from our houses (the illustration for this posting), which has been there 'ever since,' and is now mowed around as the oldfield roadside has been converted to a rough lawn. On the St-Lawrence Street/Buker Road corner lot there's a clump around a 21 cm DBH White Spruce, which presumably dates from the planting of the Spruce in the early 1990s, a stand in the lawn around a cluster of Siberian Elm trees at the edge of a septic mound, probably from not long after the construction of the mound around 1981; a roadside 5 x 7 metre stand between a small Common Lilac and a shading Sugar Maple, a ring of plants around a big Manitoba Maple, and a 5 metre stand perpendicular to the street along the boundary between lots back from a big Manitoba Maple stump.

On the Mill Street quarter of our 'doing the streets' survey, there is a 13 m stand of Daylilies along an overgrown hedge of White Cedar that's recently been cleared of lower branches to give the Daylilies room to thrive. Buker Road SW of the village (the route of the 21 July survey) is a festival of stands: a 13 metre roadside stand in a lawn, a narrow 10 metre stand along a lawn at a Cathartic (=Common) Buckthorn thicket, a 5 metre stand around a Buckthorn clump, 45 metres of scattered Daylilies in an unmowed verge along a lawn, with Narrow-leaved Plantain the conspicuous co-dominant herb, a 2 metre clump at a new driveway with Buckthorn and Honeysuckle sprouting where shrubs had been cut down, and a not-dense 2 metre clump under an Apple tree.

Other settlements are perhaps not as well daylilied, though Dominionville has "stands all through the village," in Kemptville four stands are visible from the daughter's house, 10 stands can be counted in driving County Road 18 through Oxford Mills, and there are plants sparsely along the 'wild' north side of County Road 20 through Oxford Station. In Garretton, Daylilies are dominant along the roadside of County Road 18 along Kemptville Creek.

So it seems that Orange Daylilies are now mostly around homesites or in villages, that the stands have spread only a few metres over the course of decades, and the wild plants seem to be either isolated dense stands, or strung out along roadsides where road maintenance grading could have spread plants or tubers. We saw only one field with Daylilies through it, and no cases where scattered roadside plants had become a solid stand.


  1. Appendix 1 - Matt Keevil comments on the equality of means and standard deviations of the distances between stands: "Well, I'm not sure what to make of equal mean and standard deviation except that I can't immediately come up with a reason that it shouldn't occur in a number of distributions. In the Poisson distribution the mean and the variance are defined as equivalent. I think that mechanistically a good candidate distribution is the exponential. The exponential is a case of the Gamma Distribution and describes inter-event times. Distance and time can be equivalent if we imagine taking observations while traveling at a constant speed so the rate at which Day Lilies go by might well be best described using an exponential distribution. For the exponential there is only one parameter, the rate parameter λ. The mean is 1/λ and the variance is 1/λ^2. I calculated lambda independently for each of the mean and sd for the three surveys. If the exponential is a good fit than presumably both calculated λ values based on the mean and the variance should be similar for each survey. This is indeed the case for 21 July (0.000979 vs 0.000895) and 24 July (0.000395 vs 0.000392). 2 August is a little more different, but not wildly dissimilar (3.24E-5 vs 5.56E-5) and in that case the sample size is lower. Whether these values are close enough is tough to say but I suppose a goodness of fit test could be applied to check for significant departures. The λ values are small because (I think) they represent the probability of observing lilies per m traveled."


Post a Comment

Popular posts from this blog

Mudpuppy Night in Oxford Mills

Arrival of Juvenile Leopard Frogs