Sunday, May 8, 2016

The Species

What is a Daylily?
Part 2
The Species of the genus Hemerocallis 

The phylogenetic relationships of the Hemerocallis species is an interesting if sometimes confusing subject.  This old genus of Asparagales monocots has been in proximity to humans for some tens of thousands of years, modern man having been in Asia since at least 50.000 y. a., and Hemerocallis appearing to have been there for more than 25 million years. 

Hemerocallis has been an agricultural crop in much of Asia for a couple of millennia, if not longer. For me, this long term potential for intermingling of captive-selected, feral and true, wild-species materials leads to potential admixing and selection for type that might not happen in fully wild conditions with no human intervention. 

Many Hemerocallis species can easily be transported long distances with high survivability and adaptability. This is especially true of some of the fulva group, but also for several of the yellow species. This leads me to suspect that original species ranges have long since been blurred. For that reason, I prefer to look at the species of the genus Hemerocallis in the most minimalistic terms. It is important to stress that this is not a scientific thesis, but rather a collection of my thoughts and observations about the genus Hemerocallis and the pre-domestication materials that we might refer to as 'the species'. 

I am making no attempt here to lay out a new understanding of the speciation of the Hemerocallis. I am intentionally using the most minimalistic way of understanding the two categories that the species fall within - "orange" and "yellow". This allows me to speak in more general terms and leave specifics to those with the expertise to pursue such matters.

While I may discuss where I think species boundaries may fall in some instances, this will not be my primary concern in this article. Much more work needs to be done on the phylogeny of the Hemerocallis before we can make any certain statements, but based upon current models, we see two major divisions early on in the Hemerocallis - orange and yellow. Another split seems to consistently lie within the yellows creating two major sections of yellow flowered types, but from which no clear patterns of phenotype emerge to me, as yet, with those said to clade together. I will give more consideration to possible species with the yellows than I will with the fulva group, as I think the only speciation within the fulva group is toward H. sempervirens and H. aurantiaca (or away from them). Otherwise, I think the fulva complex is composed of regional variations and garden selections - forma - not subspecies or species within an over-all 'fulva umbrella'.

I do acknowledge that there are likely to be several species amongst the yellow branch of the Hemerocallis genus. However, I think the fulva clones (or forma) all comprise one species with many regional and garden variations, clones or forma (forms). I do think that H. sempervirens and H aurantiaca are offshoots of the fulva complex, though they also could represent ancestral, evergreen types. I simply do not know, but would encourage research to elucidate this relationship more fully. Perhaps these are in the process of speciating or perhaps they represent older, evergreen ancestral forms. It is interesting to ponder. 

I do suspect that sempervirens and and/or aurantiac are of fulva origin, one way or the other, and they group with the fulva in the clades I discussed in part 1 of this series - What is a Daylily? The Genus Hemerocallis. As I am not a phylogeneticist, I am making no attempt to reassess the current species designations, but it seems some reconsideration may be necessary in the future. That however will be up the the phylogeneticists. I will use the species names as presented within the paper from which it derives, when discussing species. 

That does not mean that a plant labeled as one species is the same exact clone as that used in another project and called the same species: there is much regional variation between species, and there are numerous forms of the species in commerce and agriculture, as well as the variations that occur naturally within the wild populations of the species. I must stress that I find the entire species structure of the Hemerocallis to be tenuous, formed on the old field-identification methods of the last three-centuries, as it is. 

Now, with genetic analysis we can make much more certain comparisons, and no longer need to compare visual presentation in an attempt to determine relation. We can begin to look beyond form and understand the actual genetic relationships. With that said though, much more extensive research is needed to unlock a broad-based phylogenetic analysis. What we have now in the papers listed at the end of Part 1 is a good beginning to point us in the direction of further research and give us a starting point to piece together this complex web of relationships.

In the papers I cite at the end of Part 1 concerning the phylogenetic trees of the Hemerocallis species, there tended to be two major splits - orange and yellow - the fulvas and all the yellows comprising two different major clades. Then within the yellows, there are typically two major breaks with several species in each branch. 

Tomkins showed H. citrina vespertina to be at the base of the separation of the yellow and fulva types. The vespertina species may be suggested then to be older than other yellows and may be ancestral to all of them, as McGarty suggests in his analysis Phylogenetics, DNA, Classification And the
Genus Hemerocallis. Further data may refute this, but the phylogeny that the Tomkins tree presents is certainly suggestive and interesting. 

Some of the yellow forms are nocturnal, but certainly many are diurnal. The plant of H. citrina vespertina that I grow is nocturnal. My plant derived from Joseph Haliner. I can say nothing as to how this plant would compare to the plant in the Tomkins study.

My plant labeled H. vespertina, my accessions notes say: 
mid-July, 4" Flower, 64" Ht, dip, fragrant, 30 bud count

I have seen much higher bud counts, and scapes taller than 64". It is a massive plant with presence, though the tall scapes do not hold up perfectly and need to lean against a fence or other tall, sturdy plant. It likely grows in tall mixed field settings and woodland borders in the wild. Its leaning scapes help to dissipate its seeds further afield. We often grow this type by tall orienpet lilium to provide trellises to the scapes. In breeding, crosses to modern cultivars with strong scapes gives offspring with strong scapes, though usually shorter than the vespertina scapes. Vespertina adds branching and bud count to the offspring. I have counted as many at seven and on rare cases, up to eleven branches on vespertina.

I have no way of analyzing whether vespertina is ancestral to all the yellows or not, but it suggests fertile ground for future research. It is an interesting finding and may be a place for future researchers to apply some focus. I hope to see more in the future on the relationship of this clone to the other Hemerocallis. In the meantime, it is an interesting garden subject and well worth growing, in my experience.

The next major split in Tomkins would suggest that H. citrina is older than and possibly ancestral to H. minor, H. hakuunensis, H. dumortierii and H. middendorfii, while on the other side of the split is H. lilioasphodelus and it is shown to be older than and possibly ancestral to H. thunbergii and H. dumortierii var. Sieboldii. 

In Genetic and Phylogenetic Relationships of Genus Hemerocallis in Korea Using ISSR by Choi, et al., the phylogenetic tree shows fulva and fulva for. kwanso as clading with H. middendorfii, while H. dumortierii clades with H. thunbergii and H. minor. Another earlier branch from the yellow group give H. coreana and H. hongdoensis, both Korean forms that have been given species status.

In the paper Phylogenetic Relationships of the Genus Hemerocallis in Korea using rps16-trnK Sequences in Chloroplast DNA by Man Kiu Huh, et al., There are three phylogenetic trees presented. The first tree is a ps16-trnK analysis using MEGA5 and shows H. fulva var. kwanso at the bottom position within the tree. From there a branch occurs that contains two forks to H. coreana and H. aurantiaca, while another branch leads to all the yellows - H. minor, H. littorea, H. thunbergii, H. dumortierii var. esculenta and H. dumortierii. H. coreana and H. aurantiaca usually show the dark 'gold' coloring that is similar to the center petal pigmentation of many fulva types - rich in carotenoid pigments.

In the second phylogenetic tree - rps16-trnK analysis using PAUP 4b10 - the divisions occur in the same manner as in the first tree from figure 1. 

When the third tree is formed using PAUP 4b10, we see a shift of the layering of some of the forms involved. While the yellows hold their form from the other two maps, the positioning of H. aurantiaca, H. coreana and H. fulva var. kwanso shift, with aurantiaca at the base, coreana representing the next branch and fulva var. kwanso next and from there the large group of yellows branch off. Even with these discrepancies, we can still see that the basic break of "orange and yellow" holds up fairly well.

In synopsis, it is my view that these relational phylogenetic trees give us a glimpse into the evolutionary history of the genus Hemerocallis. Each paper reinforces the notion of the two main groups - orange and yellow. 

The fulva complex represents the use of both carotenoid and anthocyanic factors to make the more heavily pigmented flowers which include visible eyes and midribs. Both the under layer of carotin within the petal and the top layer of anthocyanins are present in most examples. Some forms in the var. rosea category within the fulva complex show reduction of carotene within the petal and a change from the typical orange pigment to a visually pink pigment in the petals. Other forms show intensification of red, while other forms show particularly bright and clean orange tones. Yet others have a purple overlay and present a brownish, fulvous coloring. Fulva complex forma are almost certainly responsible for the majority of anthocyanic colors and patterns seen in they highly-selected hybrid garden daylilies.

H. fulva var. 'Europa' 

The yellow flowers show various levels of changes from the flower color of the fulva group. The most conspicuous change is that the surface anthocyanins seem to be lost and so the under pigments in the middle of the tepals is highlighted, becoming the "coloring" of the flower, the exact tone depending on the combination, presence and/or absence of carotenes and other yellow/gold pigments. 

These flowers appear 'self colored' or solid in color and range from darker, orange golds (H. middendorfii, H. coreana, H. hakuunensis) to medium yellow gold (H. dumortierii) to yellow (H. minor, H. citrina) to light yellow (H. vespertina and H. citrina forms) to very pale yellow (H. citrina 'Baroni', other H. citrina clones). It seems to me that the range of yellows may have to do with the removal of carotenoid pigments found in the tepals that create the deep orange/gold/yellow coloring of forms such as H. dumortieri or H. middendorfii and underly most fulva forms, creating a rich pigmented area under the upper anthocyanic layer within the main body of the tepals in most fulva forma. 

As these pigments are knocked out, the lighter-appearing yellow pigments are expressed without any interference - sort of a natural "color clarification" process. The palest forms of citrina and vespertina would then represent the most reduced pigmentation amongst this group.

H. dumortierii with its golden yellow flowers.

H. citrina - one of four clones obtained from Joseph Haliner. This one shows medium yellow flowers.

H. vespertina first year of flower at about 4' tall. In the second year, this plant produced scapes over six feet tall. The flower is a lemon yellow with a hint of creaminess. Hemerocallis vespertina is my favorite of all the yellow types.

Another form of H. citrina from Joseph Haliner. This one is much larger, more trumpet shaped and extreme nocturnal. The flower is much lighter and fades to an even paler yellow than in the picture above. Picture at sunset.

It seems to me then that the yellow flowers represent a loss of pigments as compared with the fulva types. It is a common phenomena to find simple knock-out genes that stop the production of a pigment early in the developmental chain. I suspect that a major evolution from the pigmented fulva types to the yellow flowers is the shift from obviously pigmented to subtly pigmented, and may represent an evolutionary shift from daytime orange ancestors to nighttime palest yellow forms as one of the major splits within the genus. 

Many pollinators, both diurnal and nocturnal, see within the ultraviolet range. Photos I have seen of self-yellow daylily species flowers under ultraviolet lighting show intense eyes and midribs, as strong as in any fulva, but in a different range of colors tending toward blue/purple as the ultraviolet lighting reveals the pattern. I do not know if there are any forms of the yellows that do not show eyes under ultraviolet lighting, but I think that is a question someone should be looking into and would make an interesting study. 

I highly recommend taking a look.

This link takes you to the blog Photography of the Invisible World and specifically to the blog post 

Hemerocallis (Day Lilly): human vision vs simulated bee vision; reflected UV ultraviolet photography

Another excellent photo of yellow Hemerocallis under ultraviolet light can be found HERE. This link leads to the article: A deadly passion: moths and their attraction to artificial light- July 27, 2015by Eric Dillalogue. This article discusses the nocturnal pollinators - moths.

If the yellow Hemerocallis all have ultraviolet patterning, then we may have to think of all daylilies as "eyed", with those that appear non-eyed maybe just representing a knock-out gene that removes the pattern from human-visible anthocyanins, causing the same pattern to be created in some other way, but invisible to animals within our visual range. We are not historically their pollinators, so they don't really have to flash their eyes at us, after all. 

Ultraviolet spectrum lighting can reveal the presence of the same eyes and even midribs on the 'self yellow' flowers, so is the pigment anthocyanin responsible? Or are we seeing some other effect? Has the anthocyanin simply been blocked/removed, or has it been only been changed in some way? Further, the extreme change from heavily carotenoid "gold" to medium yellow and then on to light yellow may also represent loss or change of pigments. Here we may be seeing a series of knock-out genes that make flowers lighter and lighter by knocking out the production of carotenoid pigments. Building on reduction of both anthocyanin (at least out of the human-visual range) and carotenoids may well be the pathway to creating the palest flowers, what we might call "clarification factors" in the flower-breeding hobby that give rise to many of the colors seen in domestic hybrid lines that are the furthest departures from the ancestral sources.

So, what advantage could the lighter flower hold for survival? Another area where the two types split is periodicity of the flower. While none last more than a day, there is the distinct split between diurnal and nocturnal types. All of the nocturnal types are yellow and most of them are on the paler end of the color spectrum, though not all the yellows are nocturnal and both the diurnal and nocturnal forms can show the ultraviolet eye pattern. The fulvas are diurnal for the most part. If H. vespertina is the oldest of the types as suggested by Tomkins, then the major break would be between diurnal high-pigment flowers and nocturnal low-pigment flowers, each utilizing the pigmentation that serves to best attract pollinators in the light conditions in which the bloom opens.

However, at this time, that is pure conjecture. We do not know if vespertina was the ancestral style of the yellows, nor if the yellows have interbred with the oranges at times in the past giving rise to hybrid populations of orange and yellow that may still be with us as some of the current species. Clearly hybrids of yellows occurred - citrina is often self-sterile, for instance and Tomkins lists it as base to several other species. Yellow could have emerged without nocturnal behavior, certainly. Perhaps several types of yellow arose over time and some of those became nocturnal (and paler - less pigmented) over time as they colonized new niches. I hope to see much more lab work done of the Hemerocallis genus.

The evolutionary point of the pale nocturnal yellow flowers with their stunning array of ultraviolet patterning is to attract the attention of night pollinators - generally moths. They also exude a sweet smell to make themselves more attractive. The day-blooming fulva do not tend to have any fragrance and rely on a completely different group of pollinators. While we don't think of fulva being fertile due to most of us having H. fulva var. 'Europa' as our reference point, the fulva complex contains many fertile forms and is an ancient breeding population, not just one big, ancient vegetative clone. The many forms of fulva that we know of occur precisely because the fulvas have been a breeding population for a long time, generating new and novel combinations. 

Mixed fulva seedlings

All the phylogenetic trees seem to indicate that the split between the yellows and oranges is quite old, and such a split to allow the exploitation of two different environmental niches would allow genetic drift from the ancestral base by limiting the possibility of inter-breeding between night and day populations including the fulva group, except perhaps in rare circumstances. This is the very type of thing that drives speciation, as the restriction of gene flow would allow for the emergence of distinct populations. I am not saying this is how it happened, only that it makes sense from an evolutionary perspective and is a well-known driver of diversification in many species. The key to becoming a species is to keep your gene pool closed. The orange fulva group and the nocturnal yellows seem to have been fairly good at separating themselves through this process. Diurnal yellows though may naturally cross with orange types when bloom occurs in close enough proximity in space and time.

However, keeping the gene pool closed is obviously not something that has been as successful within the yellow group. Some of these types appear to be self-incompatible and thus must outcross with a different form of their own species or another species or species hybrid in order to produce viable seeds. One of the most important comments in Tomkins is the last sentence of the last paragraph of the paper, "citrina is self-incompatible and any variant arising from it would have to be obtained from an outcross. Thus, these early species variants may have arisen via cross-pollination or they may represent distinctly different genetic types". In other words, we don't really know, but we are beginning to get a window into the process.

What does seem to be clear is that 1.) there are some old, base types within the yellow category that likely represent species, 2.) all the yellows are closely related and likely share a common ancestor that broke off from the orange fulva group a very long time ago, and 3.) the actual web of relationships - what are the old base forms and what are hybrids of those or even of yellows and oranges - is not fully determined but seems to offer fertile ground for further research.

In addition to the change from orange to yellow, there are other interesting changes from the majority of the fulva group to the yellow group. One major change is that the fulva category all are rhizomatous to one extent or another. This is a good survival strategy, as it allows the plant more 'mobility' within the environment than clump formers who must depend on seed spreading to move about the environment. Rhizomatous growth is often a mark of a generalist species. Clump formers are more place-dependent, relying on other forces to distribute their seeds in order to move from place to place over the generations. Both strategies work and rhizomatous growth does not rule out sexual reproduction. However, all yellow types are not clump-forming, with some of the yellow species able to run, though none as vigorously as the more robust fulva types.

So we can assess several major diagnostic traits found within the genus Hemerocallis. They are: 1.) the presence or absence of water-soluble anthocyanic pigment in the upper, surface layers of the tepals, which is visible to the naked human eye, 2.) rhizomatous growth versus clump-forming, and 3.) diurnal versus nocturnal flowering. 

The first trait (anthocyanic pigment) is fundamental to the split between the 'orange' fulva group and the 'yellow' group. Rhizomatous growth is most common in the fulva complex and its relatives such as H. sempervirens and H. aurantiaca, but is also seen in some of the yellows such as H. lilioasphodelus. We can then say that most yellows also tend toward clump forming rather than spreading, rhizomatous growth. The third trait, nocturnal and diurnal flowering also tends to split along the color lines with the fulva group tending to be diurnal. Yellows are both diurnal and nocturnal, but we can say the majority of nocturnal forms are found amongst the yellow group.
H. vespertina clump

These three trait sets alone represent a wide range of diversity and show several levels of adaptation and response to environmental, geological and evolutionary imperatives. They show one group adapting into available niches over a long period of time and changing in response to changing environments throughout their long history in Asia. This diversity has made the Hemerocallis genus fairly generalist and adaptable. The few that might now be considered specialists are the exception.

I have not yet discussed foliage habits in regards to the species. I would like to look at foliage a bit now. As I discussed in Part 1, the Asparagales that the Hemerocallis arose from were tropical plants in a warm, tropical forest planet. These would have had little need for any genetic ability to withstand cold, temperate climate and would have probably tended toward an evergreen/ever-growing growth much as we see in current 'evergreen' daylily cultivars and Hemerocallis species such as H. aurantiaca and H. sempervirens, as well as Hemerocallis relatives such as orchids or agave.  The change in environment that brought about the formation of the Hemerocallis would involve adaptations to survive cold, as the entire evolutionary history of the Hemerocallis is intimately tied to the long plunge into the ice ages. 

When I look at the fulva complex, most of the plants are actually some form of "evergreen" and are not fully deciduous unless they are frozen off. That is, they will continue to grow to some extent if weather conditions are good, though they do not show the continuous growth of sempervirens or aurantiaca. They are 'conditional' in that they have to reach a certain level of cold to have their leaves die and then stop growing, appearing to be dormant and often forming resting buds in the north, but requiring little in the way of triggering to return to growth. Any dormancy in these Hemerocallis seems to be very dependent upon the environment. 

Northern growers regularly tell me that this or that fulva is a "hard dormant", while southern growers routinely tell me they range from "semi-evergreen to evergreen", depending on the form. The more evergreen and tender forms like sempervirens and aurantiaca will often suffer winter damage, with sempervirens being the least tender of the two. It is slow growing in the north, but it does survive without die-back, while aurantiaca will frequently show die-back from freezes.

The other fulva types tend to be much less bothered by cold, but are not necessarily frost tolerant. That is, their foliage may be killed or damaged by late freezes, but tend to recover quickly. Bloom may be effected by late freezes, as many of the fulvas bloom fairly early in the season. There are some notable late blooming examples though, such as Hankow and sempervirens, and their flowering is not typically effected by late freezes.

The foliage of many the yellows tends to follow suit with the fulva types, being conditional and responding to the environment. Much like fulva forms, the citrina, vespertina, altissima types I have grown were conditional in their dormancy, remaining in growth up to certain temperatures and then once frozen off, going into something like a dormancy, while in milder years remaining in an evergreen state and not going dormant. However, many growers in the south tell me these are "evergreen or semi-evergreen", while Northern growers often call them "dormants". 

The large dark green foliage of H. citrina vespertina and H. citrina, back right and back left. Domestic cultivar in foreground right.

I have witnessed both types of foliage behavior from the same species clone depending on climactic and weather conditions. However, here H. dumortieri is fully deciduous every year and forms a resting bud under the past year's dead foliage that persists through the winter and comes out again the next spring with little variation, generally appearing in March/April. While this looks like a true dormancy, it may not be. It could simply be an artifact of environment. 

The subject of foliage is something I wish to leave to the experts to decide. It is a complicated subject and is best left to those who are trained to make educated assessments. What I would like to look at is the utility of developing coping mechanisms if you are a tropical forest lilioid grass moving into a long, cold age of the earth that will proceed in fits and starts for several million years. Your coping strategies must revolve around ways to handle the cold. You can let your leaves die and maintain the roots and crown. If you funnel resources into making the roots frost resistant and thus able to survive freezing weather, and give up on the leaves, you can preserve energy and persist until better conditions for growth.

The development of frost tolerance would have been driven quite simply by intermittent cold periods killing out those individuals that could not survive. The colder the temperature fell, the more tender material that would be removed, except in refugia where very ancient lineages of the original ever-growing/evergreen type could potentially persist. This selective pressure would have seen the survival of those most able to handle the cold, a natural selection toward cold-hardiness, both through coping mechanisms and evolutionary changes. One excellent way to do that was to rest in a dormant-like state from which growth could be reinitiated quickly and loss through extreme freezing is only a setback; thus the main plant can persist and regrow rapidly.

Over time, changes could allow for greater frost tolerance in some lines, as well as a variety of "dormancy"/deciduous triggers to initiate rest, whatever that might actually be. I look forward to further research and invite all research into daylily foliage behaviors. I do think there are more extreme versions of dormancy in the deciduous daylilies that should be looked into. These may not represent true, classical dormancy, but may still be fully replicable and reliable traits that are expressed over wide environments. Only time and more research will tell.

The age of the Hemerocallis genus and the period of its evolution into the modern world leads me to believe that there are many possibilities, and likely more than one set of mutations, involved in foliage type and behavior. It seems to make the greatest amount of evolutionary sense to me to maintain the ever-growing traits of earlier ancestors while also being able to survive cold and freeze, mimicking dormancy for necessary periods, in some instances even going so far as to form something like resting buds, and still be able to grow quickly when the right environmental factors are present. 

To me, this conditional, ever-growing type that can stop or slow growth based upon environmental conditions and that can survive cold and freeze, loosing foliage through deciduous action or freeze damage, and then resting until the proper trigger-level of light, water and/or warmth returns to initiate growth is the most natural path an ever-growing/evergreen would take to survive in a cold world. In this way, it has mechanisms to cope with both environments when they are presented and is as such a generalist, being able to exploit a variety of major environmental challenges.

H. sempervirens

All the forms of dormancy and/or winter/cold coping strategies that we may witness in the Hemerocallis would be an evolutionary response to the environmental changes that drove them to emerge from their early Paleogene Asparagales ancestors - increasing cold and the responses that would allow those in the path of that cold to survive, where access to southern tropical and semi-tropical refugia were not available. It is unknown if there are any remnant populations that represent primitive forms. Sempervirens and aurantiaca may be late branches off of fulva that are adaptations to a warm climate and only resemble older, root forms in their evergreen growth habit, but even sempervirens will go into a partial dormancy and appear dormant in very cold frozen conditions. 

The Hemerocallis species have become as we see them through a long period of climactic change. They have made adaptations to exploit various ecological niches as the great tropical forest of the early Paleogene gave way to the drier and less-forested world of the ice ages. The landmass of Asia, showing less cooling than other parts of the earth, has provided a perfect environment for both temperate and semi-tropical forms of Hemerocallis to persist and thrive over a very long geological age. The earliest splits of type are very old and seem to revolve around flower pigmentation and time of flowering, as well as clumping or spreading growth-habit, while intermediate types may represent hybridization between any of the root forms. 

Once man became able to intervene in the destiny of plants though, the nature of the evolution of Hemerocallis would inevitably change, with both the production of food and the desire to collect exotic flowers being ancient pursuits, especially in Asia. I suspect that domestic selection has given us some of the more interesting things we see in the fulva group, while natural variations in ancient lines could have naturally produced doubled flowers or naturally more pink or red flowered individuals. Any effort by early Asian farmers or plant collectors to bring these types together could have resulted in lines where selection and improvement could have occur. 

The same is true for the yellow types, especially the citrina type which has been in domestic production for centuries and is recognized to occur in many forms in various regions. Improved clones are even offered in Chinese agriculture. Other species of yellows must have undergone a great deal of domestic selection over the centuries as well, especially where they are used in agriculture.

Next time we will look at the domestication of the daylily in ancient times and how that lays the groundwork for the modern garden daylily.

H. fulva var. 'Korean' (Seoul University - Apps)