One of the most common vegetative propagation mechanisms for cultivated waterlilies in our water gardens is generally called “viviparism”-- which generally refers to a living being that produces live offspring, and in botany to those plants which produce bulbs or seeds that germinate while still attached to the parent plant-- and surely it is close to our experience since a good number of these plants display such behavior as a reproductive strategy. But we must keep in mind that these waterlilies, mostly hybrids or cultivars, inherited this feature from one or more of their ancestors. If we go back through successive generations, somewhere along the line we are bound to find the link with the species that originated this particular feature.

Among the approximately 52 known species, “viviparous” behavior is pretty rare and those that display it habitually can be counted with the fingers of one hand. In fact, only three known species today have that reproductive form as a primary, almost primary, or combined resource: Nymphaea micrantha Gill. & Perr., Nymphaea lasiophylla Mart. & Zucc., and Nymphaea prolifera Wiersema.

What can be highlighted at first sight from these species?

• All three are tropical. • Only one of them (N. micrantha) is a day-bloomer, while the other two have nocturnal anthesis. • Among recognized subgenera, N. micrantha belongs to Brachyceras while N. lasiophylla and N. prolifera belong to Hydrocallis. • Their type of viviparism:		N. micrantha
- N. micrantha with its gemmipary(1) property (meaning the development of young plants from vegetative buds borne on the parent body) generates new individuals only from its leaves, developing epiphyllous plantlets (from the Greek meaning “upon a leaf”). - N. lasiophylla and N. prolifera do it from their flowers, by means of abortive tuberiferous flowers in a process called pseudovivipary(1) by Sculthorpe in 1967 and vivipary (2) by Wiersema in 1987.		N. micrantha
		N. prolifera

Back from botany to our daily experience, as noted above there are several viviparous cultivars. Among those we consider tropical, their viviparism from the leaves corresponds, in ALL cases, to their N. micrantha descent, namely N. ‘Daubenyana’, N. ‘Panama Pacific’, N. ‘Tina’, N. ‘Lindsey Woods’ and many more.(3) These waterlilies evidence their viviparism in a very pronounced manner, since adult plants will make this mechanism their most common form of propagation.

It is altogether different when we refer to hardies, where in some this phenomenon is rare or almost unseen -- might even never be seen -- but it exists. In others it is somewhat more frequent. To summarize, we could say that those that present this feature more or less markedly are slightly or eventually viviparous and their reproduction does not depend on this particular characteristic. Here N. micrantha is not involved, since in this group of waterlilies in ALL cases there is flower -- but NOT leaf -- viviparism.

We could then feel tempted to consider that viviparous hardies have some kinship with N. lasiophylla or N. prolifera, since only these species present vegetative propagation in their floral structures as indicated above.

Is this possible? Let’s say that almost anything is possible, but it is not likely, because even though N. lasiophylla was described in 1832 there are no references to its being used in any hybridization. Furthermore, we know that they belong to another subgenus, and intersubgeneric crosses have been known and proved in more recent times (I admit, by the way, that this is a rather weak argument). Let us add that the nonexistent horticultural attraction of this species would make it unlikely for a professional hybridizer to utilize it aiming to improve future generations of cultivars. Lastly, we could consider as proof of its use that, besides its being viviparous, its night-blooming characteristic could have been transferred too -- but we do not have night-bloomers among hardy waterlilies … yet.

Similar considerations might be added by specialists pro and con, but ultimately no one can really attest or deny the N. lasiophylla descent, which is why we subscribe the UNLIKELY hypothesis given its more limited scope.

So if we discard N. lasiophylla and leave aside N. prolifera for identical reasons, besides its being unknown in a horticultural sense, we are left with nothing. So how is it possible that we have flower-viviparous hardies? 

There are some who think, among other considerations, that N. ‘Colonel A.J. Welch’ could be a non-described species that became extinct in its native or natural state. Most are led to believe, however, that Latour-Marliac (from whose business it originated, although it cannot be said that he himself hybridized it) or whoever actually made it used N. mexicana and therein might lie the answer. This makes sense, because Latour-Marliac used N. mexicana extensively and N. ‘Colonel A.J. Welch’, given its color and its flowers above the water level, surely carries its descent. Paradoxically, experts express some doubts by noting that N. mexicana is not described or known as viviparous from its flowers; it generates runners or clusters in the shape of bananas in the root area.

In agreement with the above view, in my opinion N. mexicana has caused this particular feature in viviparous hardies, including N. ‘Colonel A.J. Welch’. To substantiate this position, I offer the following evidence: 1) a scientific article published in India in 1974 which explains this (4); 2) the testimony of Grant Mitchell, known for his love of lotus and his trips to the Far East, reported extensively in our lists, who some years ago in an IWGS forum said he had observed a flower in N. mexicana with a little plantlet growing on it, and lastly 3) the experience of a founding member of WGI, the now famous hybridizer Pairat Songpanich, for having achieved the “almost impossible” blue hardy called N. ‘Siam Blue Hardy’. He has made use of N. mexicana for hybridization and says it is a flower-viviparous waterlily based on his personal experience. This feature, so far unique within the genus Nymphaea, is certainly quite elusive, it does not appear in all N. mexicana or is very weak, or else depends on very specific conditions, and that is why it has largely not been observed.

This could have been the reason why using such a precise type of N. mexicana with that particular feature to obtain N. ‘Colonel A. J. Welch’ gave its descent such a strong trait, while it did not happen with many other hybridizations past and present using the same species (N. ‘Marliacea Chromatella’, N. ‘Moorei’, N. ‘Aurora’, N. ‘Andreana’, N. ‘Yellow Sensation’, etc.)

It should be noted that the names of cultivars mentioned here displaying this feature, even if observed only once, were provided by some of those for whose help I thank at the end of this article and who responded to my queries.

There are, however, some occurrences that seem to be exceptions for the time being. Again, Pairat Songpanich tells us about some of his hybrids, all of them “odorata” rhizome types without apparent kinship with N. mexicana, that have also shown this behavior. Such are the cases of N. ‘Mayla’ x N. ‘Perry’s Fire Opal’, N. ‘Perry’s Fire Opal’ x N. ‘Sirius’, and (N. ‘Perry’s Fire Opal’ x N. ‘Sirius’) x N. ‘Sirius’.

My initial working hypothesis before writing this article was trying to prove the incidence of N. mexicana in this type of viviparism. I believe this objective has been achieved, although there are still some obscure areas which leave room for further learning or understanding about this exceptional behavior. Later its scope widened to cover all hardies, and finally it came to include all the lesser known aspects of the family.

1) Barber, C.A. (1889), “On a change of flowers to tubers in Nymphaea lotus var. monstrosa”, Annals of Botany IV (13): 105-116.

2) Mitra R.L. and K. Subramanyam (1982), “Is Nymphaea rubra Roxb. ex Andrews an apomict?”, Bulletin of the Botanical Survey of India 24: 83-86.

Perry Slocum also assigned flower viviparism to N. lotus in the description of this species in both editions of his book, An Encyclopedia of Water Lilies and Lotuses (6).

We should also add the personal comment by the renowned hybridizer Craig Presnell from Florida, who confirms the presence of flower viviparism in N. ‘Missouri’ of N. lotus descent; the report by Dr. N. Nopchai Chansilpa who attests to having seen several offspring of N. lotus with this behavior; Nan Bailey’s report from Australia on an unnamed night-blooming hybrid; and a report from my own country, Argentina, on N. lotus specifically.

This report on Anecphya is the first and only one to date. Andre Leu, an unavoidable reference and keen specialist on this subgenus, refers to this case as “an oddity of nature”. Even considering it exceptional, as it surely is, the ability to propagate in this manner indeed exists.

Before we came upon this information, such asexual propagation by abortive tuberiferous flowers appeared to be related only to subgenera around the syncarpiae group. Dr. Wiersema thus indicated in his description of N. lasiophylla: “… rare occurrences of similar structures have been reported for other syncarpous Nymphaea,” and suggested reading Moham Ram et al (4) and Mitra and Subramanyan on N. rubra.

As a statistical synthesis, we can state that we ended up with flower viviparism events in both groups (syncarpiae and apocarpiae) and in four (Nymphaea, Hydrocallis, Lotos and Anecphya) of the five or six admitted subgenera, and with leaf viviparism in one group and one subgenus (Brachyceras). As a result, each of the family’s taxonomic division or subdivision has at least one representative of viviparous behavior.

We should also point out that of the 19 waterlily names included in this report (not including N. ‘Colonel A. J. Welch’), 10 are cultivars obtained by Kirk Strawn. Considering that some of his other cultivars besides the above-mentioned have a similar genetic descent, with N. mexicana as polen parent, it would not be surprising if one of these might be added to the list in the future when such an eventual occurrence is observed and testified by a fellow water gardener.

One final comment: the most amazing reports came from the Orient, mainly on N. ‘Helvola’ and N. ‘Lily Pons’ and on “odorata” without the alleged intervention of N. mexicana. It seems this phenomenon is less observable in the West, while in Thailand with Pairat Songpanich and Dr. N. Nopchai Chansilpa, and in China with the great hybridizer and lotus and waterlily specialist Prof. Huang Guozen of the Quingdao Biological Institute (Quingdao Water Lily World), they see it as a rather common behavior.

I am grateful to Roger Bagley, Nan Bailey, Mauro Bergamo, Paula Biles, Dr. N. Nopchai Chansilpa, Alani Davis, Daniel Flouret, Huang Guozen, Don Hamburger, Harro Hieronimus, James Knock, Andre Leu, Carlos Magdalena, Asta Miklius, Walter Pagels, Craig Presnell, Andreas Protopapas, Rich Sacher, Pairat Songpanich, Martin Sturge, Charles B. Thomas and Werner Wallner for their participation. I want to express my special gratitude to my friend Dr. Jaime Raúl Bonilla Barbosa, a professional specialist in aquatic botany at the Autonomous University of the State of Morelos, Mexico, who made a technical and formal revision of this article.

References:

1) Sculthorpe, C.D. 1967. The Biology of Aquatic Vascular Plants (London: Edward Arnold) 610 p.

2) Wiersema, John H. 1987. A Monograph of Nymphaea subgenus Hydrocallis (Nymphaeaceae). Systematic Botany Monographs 16: 1-112.

3) Viviparous Waterlilies - www.victoria-adventure.org

4) Mohan Ram, H.Y., Laxmi Nayyar, Vijay (Department of Botany, University of Delhi).1974. “A Case of Reversion of Flower of Nymphaea mexicana to Vegetative Condition”, Current Science 43 (9): 291.

5) James Knock's Tribute to the Brilliance of Kirk Strawn - www.victoria-adventure.org

6) Perry D. Slocum, Peter Robinson and Frances Perry. 1996. Water Gardening, Water Lilies and Lotuses (Timber Press).