What about Psammirises,
Pseudoregelias, and Falcifolias?
These irises are all small in stature, with some points of resemblance to both bearded irises and to arils. Except for Iris humilis (formerly referred to as I. arenaria), which has an enclave population in eastern Europe, they are all native to the steppes and mountains of central Asia. The psammirises (which means "sand irises") include I. humilis, I. bloudowii, I. potaninii, and others. The best known examples of the pseudoregelia group are I. hookeriana and I. kemaonensis, and I. facifolia is (nearly) alone in a grouping by itself. These irises do have beards; their foliage is grassy, narrow, and wiry, and their rhizomes small and tightly clumped. They have miniscule arils on their seeds. I shall mostly focus on the psammirises, as this is the only group that was ever frequently grown in gardens or used in hybridizing. I have grown and bloomed I. potaninii (pictured at right) from a commercial source, and germinated seeds of I. humilis that have yet to grow and bloom here.
Here is the current scientific understanding of their evolutionary relationship to other irises, based on molecular DNA data:
In reading such a diagram, an analogy with a human family tree may be helpful. The two aril groups, oncocyclus and Regelia, are siblings, more closely related to each other than to any of the other groups. The bearded irises (both the dwarf bearded species and the tall bearded ones) are their "first cousins". The psammirises are one step further removed, "second cousins", as it were, to both the arils and the bearded irises. The next step further out we find "third cousins" I. dichotoma and I. domestica, which you may recognize as the parents of the popular garden plant I. X norrisii. Other groups in the genus are even more distant relations.
So these irises are neither bearded irises nor arils (as the term is currently used), but are more closely related to both those groups than any other species in the genus. At one time, their relationship to arils and bearded irises was thought to be even closer. It is worth reviewing this history, as it has influenced how iris enthusiasts have treated them over the last century.
W. R. Dykes, in his monumental 1914 monograph The Genus Iris, treated each of these three groups differently. He placed the psammirises among the bearded irises, in two separate groupings, alongside the dwarf and TB species. He placed the pseudoregelias in a group by themselves, separate from oncocyclus, Regelia, and bearded irises but related to all three. I. falcifolia he simply listed as a Regelia, alongside I. stolonifera and I. korolkowii, without comment.
Dykes's placement of the psammirises among the bearded species prompted dwarf iris breeders to take an interest in them, and about twenty hybrids between them and the dwarf bearded cultivars of the time (mostly forms of I. lutescens) were created, registered, and introduced into commerce.
This was superseded by the classification of Lawrence and Randolph (1959), which was the most influential treatment for most of the English-speaking world in the mid-twentieth century. This is the system used in the essential reference The World of Irises, published by the AIS in 1980. Here is a diagram of this system:
Here you can see that the psammirises and facifolias have been placed in close relationship with the Regelias, closer to them than even the oncocyclus. The pseudoregelias remain in their own category, much as in Dykes's system. Around this same time, the iris hybridizer and nurseryman Lloyd Austin popularized the term "aril" to refer to the irises with arillate seeds, particularly the oncocyclus and Regelias that were the object of much attention at the time, but also including our three groups, based on the presence of arils on their seeds. Just as Dykes's treatment of the psammirises prompted iris lovers to regard them as dwarf bearded species, Lawrence's treatment and Austin's promotional efforts prompted people to think of them as arils, kin of the Regelias in particular. The korolkowii/humilis hybrid 'Koviar' (1970) is the sole example of a hybrid registered from aril/psammiris breeding. The Aril Society International, founded in 1957, classified all three groups as arils, under the influence of Austin's terminology and Lawrence's classification. Indeed, it would have been inconsistent to exclude the psammirises and falcifolias in particular, given their perceived closeness to the Regelias. The psammiris/dwarf hybrids were treated as regeliabreds at this time, and for that reason I have included entries for them in my Checklist of Arilbred Dwarfs and Medians.
The next major revision of the classification of the genus came with Brian Mathew's publication of Iris in 1989. He determined there was no compelling evidence to connect these irises with the Regelias in particular, and set out the following system:
This system may be seen as an expression of humility: Mathew knows that the oncocyclus, Regelia, bearded irises, and our three groups are all related to one another - somehow - but does not commit to anything more specific than that. The Aril Society took notice of the change and restricted its scope to the oncocyclus and Regelias. From this point, the psammirises, pseudoregelias, and falcifolias were no longer "arils", although they might still be called "arillate" in reference to the morphology of their seeds. It is worth noting that arils are found on the seeds of other irises, such as the Junos, that are not particularly related to oncocyclus and Regelias, and even on many plants that are not irises. They are a trait associated with seed dispersal by ants and other insects, not necessarily a marker of relatedness.
When we now look at the relationships established by modern DNA analysis, we can see a trend in the understanding of these irises over the past century or so. Initially, the temptation was strong to connect them with the more familiar bearded and Regelia irises. Dykes could matter-of-factly list I. falcifolia as a Regelia species, and all the psammirises as dwarf bearded species, right along with I. pumila, I. lutescens, I. suaveolens, and the others. Lawrence connected the psammirises and falcifolias with the Regelias, but gave them their own pigeonholes in that association. Mathew disentangled them from the Regelias. Finally, the DNA analysis places them even farther afield. We now know that the arils (oncocyclus and Regelia) are closely related to each other, and that their next nearest relations are the bearded irises, not the psammirises. The psammirises and their kin are also related to both arils and bearded irises, but more distantly. They should not be associated with either arils or beardeds in particular, but are part of a larger clan that includes all those groups.
As mentioned above, psammirises have been successfully crossed with both bearded irises and arils. I am not aware of any crosses involving the pseudoregelias or falcifolias.
No fertile seedlings have resulted from either type of cross. 'Tiny Treasure' (Hillson, 1943) has a registered parentage of 'Ylo' X I. humilis, and 'Ylo' is a humilis/lutescens hybrid. However, a chromosome count was performed and the result was not consistent with the parentage, which was likely recorded in error.
The psammiris/lutescens hybrids have a distinct charm, and some are still being grown. Early dwarf hybridizers were attracted to I. humilis and I. bloudowii for their flaring form, clear yellow color, and the habit of the faded flowers to curl up into tight corckscrews, rather than just going limp as bearded irises do. When Jay Ackerman succeeded in crossing I. humilis with I. pumila in 1956, it was a sensation in the world of dwarf iris enthusiasts. The DIS Portfolio embellished its drab mimeographed pages with unprecedented color photos, attached by staples. I think interesting results could be achieved by crossing the psammirises with modern SDBs, although hybridizers seemed to lose interest in the psammirises once the SDBs came into their own. The last bearded/psammiris hybrid ('Sunaire') was registered in 1969.
The psammirises used in early hybridizing have chromosome counts of 2n=22, although more recent counts of different collected forms more often show 2n=28. I. hookeriana (Pseudoregelia) was counted as 2n=24. I. falcifolia is 2n=18. The psammirises having the same chromosome count as the diploid Regelias added support to the conjecture that these two groups were closely related, but is now viewed as coincidental, as is I. hookeriana having the same count as diploid TBs. As discussed elsewhere on this site, it is not chromosome number per se but rather the homology (chemical and structural similarity) of the chromosomes that is important in meiotic pairing.
The compatibility of psammirises with both bearded irises and arils offers some hope to hybridizers, but fertile offspring are expected only at the tetraploid level, barring a rare unreduced gamete from a diploid hybrid. As there are no known natural tetraploids among the psammirises or other related groups, any work in this area will have to wait until someone induces tetraploidy in them artificially, or else raises thousands of seedlings in a brute-force (a la C. G. White) approach to generating accidental tetraploids or amphidiploids. Since these irises lack the intrinsic drama of the spectacular oncocyclus to inspire such laborious undertakings, it is likely that they will remain a footnote in the history of iris breeding for the foreseeable future.
If I had the resources and inclination to pursue a hybridizing program with the psammirises, I would probably put aside the task of crossing them with bearded irises or arils, which looks to be a dead end, and instead cross them amongst themselves to bring out whatever latent genetic potential is in the group and breed fertile plants at the diploid level with desirable garden qualities.
updated January 2018
Unless otherwise noted, all text and illustrations copyright Tom Waters and all photographs copyright Tom or Karen Waters. Please do not reproduce without permission.