Three topics are considered:
• General principles;
• Influence of particular tillandsia species;
• Some hybridising suggestions.
Hybridising involves a certain amount of effort; and usually a long period of time, for example 5 to 15 years, before you have a flowering plant. So some pre-planning can be very helpful. The comments in this paper are mainly based on primary crosses (F1) between species.
There is not as much information available on F2 crosses. This is due to several reasons. Some F1 hybrids appear to be infertile which limits the available pool of parents. More importantly though, as many F1 hybrids have only become readily available in the last 15 years, we have only seen a significant number of F2 hybrids recently.
This article is largely based on a paper by Mark Dimmitt (Dimmitt 1990), and input from attendees at various Bromeliad Society of Queensland Tillandsia workshops and the Australian Tillnuts’ group. It also draws heavily on an article written by myself on this topic published in the Nov-Dec 2005 edition of Bromeliaceae, and one written by John Olsen published in the 2nd Quarter 2012 of that publication.
It helps to have a clear idea as to the specific plant characteristics which you are seeking to produce. Some factors worth considering are: the hybrid’s growth habit, overall size, inflorescence, and any other specific traits you are targeting, such as its scent when flowering.
If you just cross any species which happen to be available, then the chances of you achieving an outstanding hybrid may be quite low. You may also end up duplicating a cross which has already been made.
The next step is to consider what has already been created. Many, but not all, hybrids are registered in the Bromeliad Cultivar Register (BCR): www.registry.bsi.org. The registry has details of over 1,300 tillandsia hybrids utilising over 150 species as parents. Have a look at it to see if your “dream plant” has already been created — or you may gain some more ideas as to which species you would like to use (or not!).
While the BCR registry contains a lot of useful information, it has limited information on the tillandsia species which are fragrant when they flower. Barry Stevenson and Sharon Lee recently compiled a list of these species — see https://www.facebook.com/notes/tillandsia-addicts-australia-buy-swap-sell-share/perfumed-tillandsia-species/1020508584752961/ or 'Detective' article DD08/20 on Bromeliads in Australia website.
Their list, after removing species which are now in other genera, and some additions from other sources; is: arequitae, arhiza, aurea, bandensis, brealitoensis, cacticola, caerulea, calignosa, cotagoitensis, crocata, diaguitensis, disticha, duratii, funebris, graomogolensis, humilis, ixoides, jucunda, kuehasii, linearis, mallemontii, mereliana, myosura, nuptialis, paleacea, pieranoi, prolata, purpurea, reichenbachii, remota, straminea, streptocarpa, streptophylla, tafiensis, tenebra, usneoides, xiphioides, and yuncharaensis.
A hybrid of any two of these species is highly likely to produce offspring with fragrant flowers. Some species (for example: crocata, duratii, mallemontii and usneoides) also produce at least some fragrant flowering offspring, when crossed with a non-fragrant species.
Many species vary considerably in terms of size, leaf/inflorescence colouration and other characteristics. While some species are represented by only a few clones in Australia, many others have a much wider range from which to choose. So, pick the best clones for your purpose.
You can extend the range of pollen parents available to you, by harvesting and storing pollen. It can be stored for 18 months or more, under the right conditions. So, you do not have to rely on just using pollen parents which are flowering when the seed parent’s flowers are ready for fertilising.
As a general rule, pick the larger plant of the two which you are hybridising to be the seed parent. This is because production of seed requires a large amount of energy from the plant, and the bigger the plant, the more energy it will have for seed production.
Even so, if you try to set too much seed on the plant it could die before the seed matures. As a suggestion, limit yourself to a maximum of 10 pods on any plant, and 1 or 2 on small plants such as Tillandsia ionantha. (Regular feeding with a foliar fertiliser will also help maximise the amount of seed produced). If you are using more than one pollen parent, then tag each flower with the pollen parent, to minimise the chances of a “mix up” when you are harvesting the seed.
You do need to take care that the pollen you are using to fertilise the seed parent is, in fact, the “father”, and that some other “stray” pollen does not fertilise the plant. There are various techniques which you can use to help achieve this outcome. For example, isolating the seed parent plant away from other tillandsias, and removing the seed parent plant’s anthers before they are ready to shed pollen.
It will pay to keep detailed records of your hybridising—relying on your memory is often not a good strategy! As a suggestion, record the parents used in the cross (including details of any particular clones or varieties which may have been used), pollination dates/times, fertilisation results, seed harvesting information such as harvesting dates and seed quantities obtained, and germination results.
It can take two to three generations of vegetative propagation before you can be confident as to the final characteristics for example, size and inflorescence, of a particular hybrid.
While some information is available on the traits which “dominant” parents such as T. bulbosa and T. duratii transmit to their F1 offspring, there is even less available on which species do not transmit their characteristics. Yet this information is just as important if future hybridists are not to repeat the “mistakes” of the past. So, please keep this information and let other people know about your experiences.
There are some spectacular species which are very difficult to grow in particular region. Using them as a pollen parent can be a way of “transmitting” their best characteristics into another plant (that is, a hybrid) which can be more easily grown. For example, the beautiful T. carlsoniae is difficult to grow in south-east Queensland. However, its multi-branched inflorescence which is one of its outstanding qualities, appears in the more easily grown T. Queen’s Delight (T. carlsoniae x T. chiapensis).
If a hybrid is not an improvement on its parents, then there is little to be gained by releasing it. Instead, it is best to discard them.
Naming, and registering, a hybrid is primarily the privilege (and responsibility!) of the hybridist. To help minimise confusion, where there are offspring of a given cross, with markedly different characteristics, then it is worthwhile (where practical) giving them different names, instead of just using the one name for the entire cross.
In such cases, it may be worthwhile to use a “two part name” with the first part being the same for all offspring. (An example could be: Auchenflower Magic, Auchenflower Sunrise, and Auchenflower Sunset. These would denote three distinctly different clones from the one cross).
Finally, some hybridists, due to lack of growing space, have to release seedlings from a cross before they are mature. So, the hybridist may not finish up with a plant of the best clones from the cross. If you have bought a seedling which develops into an exceptional plant, then please consider giving the hybridist the opportunity to obtain one of its pups.
Influence of particular tillandsia species
In many cases, only a few crosses have been made utilising a particular species, so the comments below should be used as a starting point, rather than as “gospel”. In some species, for example T. capitata, a wide variety of clones have been used in hybridising. So it is important to know the clone which has been used, if you decide to use T. capitata in a new cross because of the characteristics it displayed in a particular hybrid.
Further, some traits are genetically recessive, and will often not appear in F1 crosses. However, while they may not appear in F1 crosses, they may “re-emerge” in a proportion of the plants arising from an F2 cross. There is little information available on this point. The comments below relate to F1 hybrids.
aeranthos contributes its bracts and petal colouration and ease of flowering.
albertiana contributes its growth habit (including size) and petal shape/colour. It appears to be a dominant parent in many crosses.
albida “… has a caulescent habit that does not show up in hybrids with acaulescent species, but its’ very long, thin inflorescence is dominant. Its hybrids have not been winners, but one with T. streptophylla is interesting. It looks like a giant, acaulescent T. albida (or a white-leafed , non-curly T. streptophylla) with a very tall, branched, narrow inflorescence with reddish bracts and pale blue flowers…” Dimmitt (1990) p. 121.
balbisiana contributes its growth habit.
baileyi contributes its size, growth habit and inflorescence shape.
bergeri contributes its size and growth habit.
brachycaulos “…contributes soft, green leaves. It also greatly shortens or usually completely supresses the elongated spike of another parent. The bright red leaves at maturity do not come through well unless the other parent has the same trait...” Dimmitt (1990) p.121.
bulbosa “…with its bulbous base and twisted, awl-shaped leaves (it) is extremely dominant. Nearly all of its hybrids look mostly like it, to the extent that it is difficult to recognise the other parent. It tends to dwarf the inflorescence of hybrids. The trait of red upper leaves at flowering does not come through well…” Dimmitt (1990) p 121.
butzii contributes its growth habit, including size, and branched inflorescence.
capitata contributes the shape, size and colour of its inflorescence. Depending upon the clone used, it can “compress” the inflorescence of multi-branched species with which it is crossed.
caput medusae contributes its growth habit.
carlsoniae contributes its multi-branched inflorescence and growth habit.
carminea contributes its growth habit, including size.
chiapensis contributes its size and growth habit.
concolor “…transmits stiff leaves and good symmetry of the rosette. The branched spikes and their bright red colour and/or chartreuse bracts are transmitted to offspring very well. Its progeny bloom over an extended period, and the bracts stay colourful for about three months...” Dimmitt (1990) p.121.
copanensis sometimes contributes its growth habit.
crocata contributes its inflorescence’s size, shape and petal colour, as well as its flowers’ perfume.
didisticha “…greatly slows the growth of its hybrids, even with T. stricta; none are even close to flowering at four years of age…” Dimmitt (1990) p.121.
duratii. Dimmitt (1990) argued that its size was not evident in its hybrids. This statement is not universally true. For example Wonga (duratii x mallemontii) is much closer in size to duratii, than mallemontii. Dimmitt stated that duratii contributes stiff, succulent, heavily lepidote (scurfed) leaves to its hybrids. This is not always true. For example, neither Wonga nor Goomong (duratii x stricta) have succulent-like leaves. It often (but not always) contributes its inflorescence size/shape and flowers’ perfume.
edithae does not appear to be a dominant parent in its hybrids.
ehlersiana contributes its silver foliage, growth habit and inflorescence. Its hybrids grow slowly.
exserta typically contributes the size, shape and colour of its inflorescence.
fasicualata often contributes its growth habit, multi-branched inflorescence, and relatively vigorous “pupping” traits.
flabellata contributes its size, growth habit, leaf colour and inflorescence size/shape.
gardneri contributes its “soft” leaves. Its inflorescence colour is often not passed on.
geminiflora contributes its size and branched inflorescence.
intermedia contributes its elongated form, and growth habit of a few leaves, to most of its hybrids.
ionantha “…dwarfs the size of its hybrids and greatly shortens or more often eliminates the elongated spike of the other parent. The red leaves at maturity do not come through, unless the other parent had the same trait…” Dimmitt (1990) p.121. However, the cultivar ‘Druid’, does not appear to transmit its yellow/white inflorescence/leaf colour to its offspring.
ixioides “…contributes stiff lepidote leaves. The green flower bract colour is very dominant. The yellow flower colour is transmitted if the other parent has white flowers. Crossed with blue flowers, the result is dirty brown flowers i.e. with T. stricta…” Dimmitt (1990) p.121.
jalisco-monticola contributes its growth habit.
kautskyi contributes its soft leaves and large trichomes, plant shape and bract colour.
latifolia contributes its inflorescence and growth habit. (Clone selection is important with this species, as it varies widely in size and growth habit)
mallemontii contributes its growth habit, but not necessarily its size.
matudae contributes its growth habit and inflorescence.
mauryana contributes its inflorescence and growth habit.
neglecta (‘red form’) contributes its leaf colouration and growth habit.
novakii contributes its inflorescence.
pseudobaileyi “..like T. bulbosa, is heavily dominant. Hybrids have bulbous bases, twisted awl-shaped leaves, and frequently faint striations on the foliage. It contributes a large, branched inflorescence...” Dimmitt (1990) p. 122.
recurvifolia (syn. meridonalis) “…contributes stiff, lepidote leaves and fairly vigorous growth, though not as fast as T. stricta. The white flower colour dilutes the colour of the other parent in the progeny…” Dimmitt (1990) p.122.
The orange bracts of T. meridonalis v. subsecundifolia are usually not transmitted to its progeny.
rothii contributes the shape and colouration of its inflorescence and the red “hue” which its leaves assume at flowering—for example see Belli (rothii x exserta). It also contributes its slow rate of growth.
rotundata contributes its growth habit and inflorescence.
schiediana often contributes its “two tone” petal colour to its progeny’s flowers, as well as its growth habit.
seleriana contributes its size and growth habit.
sprengeliana contributes its soft leaves and large trichomes, plant shape and bract colour.
streptocarpa contributes its growth habit (but not necessarily its size), inflorescence shape/size and petal colours.
streptophylla “…transmits its large size and very large well-branched spike to almost all of its hybrids. Some hybrid progeny also have twisted leaves…” Dimmitt (1990) p. 122.
stricta “…contributes its rapid growth more than anything else, including its short-lived blooming habit. The pink flower bracts and blue flower colour are greatly suppressed when crossing with green-bracted or white flowered species…” Dimmitt (1990) p.122. Further, T. stricta often, but not always, transmits its growth habit and size to its progeny.
sucrei contributes its growth habit and inflorescence.
tenuifolia contributes its caulescent growth habit.
tectorum unfortunately, its heavily lepidote leaves are not transmitted to its progeny.
tricolor contributes its growth habit.
usneoides contributes its growth habit and size.
xerographica contributes its growth habit (including wide, greyish, succulent-looking leaves) and multi-branched inflorescence. This statement also applies to the form of T. xerographica formerly known as T. tomasselli.
xiphioides contributes its slow growth rate and dislike for wet conditions. Its large flower size is often not transmitted to its progeny.
Some hybridising suggestions
Much depends upon the type of hybrid you are seeking to produce — everyone has their own preferences!
My approach would be:
• Establish the characteristics for example, size, growth habit and inflorescence shape/size/colour, of the hybrid which you would like to produce (while taking into account the third dot point below);
• Do not repeat existing crosses unless you have some exceptional clones of the relevant species;
• Choose as one parent a species that has consistently produced hybrids which are keenly sought after by tillandsia enthusiasts and is easy to grow in your area. Examples of smaller-growing species are: T. chiapensis and T. geminiflora, while larger ones are T. fasiculata and T. streptophylla. (T. xerographica is also a great parent, but its progeny can be slow-growing, so at my age it is not a great choice as I would like to see the progeny flower while I can still enjoy them!);
• For the other parent, select from species which are easy to grow in your area and which, when combined with your other choice (see last dot point), could feasibly produce your desired hybrid;
• Use the best clones which you can find for each parent.
I think some interesting hybrids could be:
• T. chiapensis crossed with: atroviridipetala, cacticola, ehlersiana, esseriana (red form), latifolia var. leucophylla, zacapensis.
• T. fasiculata crossed with: barthlottii, latifolia var. major, socialis, zacapensis.
• T. geminiflora crossed with: atroviridipetala, esseriana (red form), mauryana, sprengeliana.
• T. streptophylla crossed with: barthlottii, latifolia var. major, zacapensis.
Dimmitt, M.A. (1990) Additional notes on breeding superior tillandsias BSI Journal v. 40 (3) 118-123.