The Bromeliad Society of Australia Incorporated.
Affiliated with The Bromeliad Society International(U.S.A.)

Some Amazing Bromeliads, Part Four.
By Derrick. J. Rowe.

Arboreal Ant-house plants revisited.

T. baileyi, T. balbisiana, T. bulbosa, T. butzii, T. caput-medusae, T. paucifolia, T. pseudobaileyi, and T. streptophylla are the species already mentioned in article two of this series as having enlarged pseudobulbous bases with hollows that field observations confirm regularly house ant-colonies. (See references there.) In addition to these eight species, I have added T. seleriana due to a field report that Derek Butcher kindly advised (Rauh 1970.) and because Benzing 2000 also claims it is an ant-house plant.

That T. seleriana is a confirmed myrmecophyte is not at all surprising. It has such a large, rotund pseudobulbous base with many cavities, that among Tillandsia species it is probably the one most similar in general outline to tuberous myrmecophytes such as Australia’s Myrmecodia beccarii (other major differences aside).

There are other Tillandsia species with pseudobulbous bases that are also probable myrmecophytes but I have yet to find field observations confirming this in the literature. They are T. arizajuliae, T. diguetii, T. intermedia and T. pruinosa all of which are placed in sub genus Tillandsia making a grand total of thirteen species. Nevertheless, just to complicate the situation, five species that according to their DNA very probably belong in this ‘pseudobulbous’ clade have reverted to no longer having a pseudobulb. They are T. achyrostachys, T. concolor, T. ionantha, T. juncea, and T. palmasolana. For example T. achyrostachys is a sister species to T. pseudobaileyi, while T. ionantha is sister to T. bulbosa. In other words they are pairs of closest living relatives species-wise but only one of each pair has what was considered to be a synapomorphic pseudobulb. A synapomorphy is an evolved feature shared by all future descendents; hence it should define all future branching in a cladogram – a diagram of an evolutionary-tree. All descendents in theory should have the same synapomorphic feature but as we can see, nature seldom fits neatly into human pigeon holes. It also shows how the use of comparative morphology without the evidences emerging from genetics and other modern sciences can be very misleading.

Two other pseudobulbous species are currently placed in subgenus Allardtia which possibly indicates that the pseudobulbous habit evolved at least twice within Tillandsia but the correct phylogeny is still unclear. They are T. disticha and T. ehlersiana. (Chew et al. 2010.) Werner Rauh 1970 notes that T. disticha is a transitional form, “the bases of the scoop-like leaf sheaths are succulent, have water-holding tissue and lie tightly packed together; the non- succulent upper sheath sections, on the other hand, form hollows inhabited by ants.”

T. flexuosa co-occurs with the ant-house species Tillandsia baileyi, T. balbisiana, T. bulbosa and T. streptophylla in a study site in the State of Quintana Roo, Mexico, yet in this study it was the only species found not to contain ants. (Olmsted and Dejean 1987.) Although not considered a member of the pseudobulbous clade by Dr Chew’s team it has a ‘bulbous’ base with large water-tight chambers and it usually grows high in tree canopies; a site typical of most sunlight-preferring myrmecophytes and their ants. Benzing 1990 however states that this species sometimes contains ants.

Aechmea bracteata, A. brevicollis, and Brocchinia acuminata are all confirmed ant-house plants that have been mentioned in a previous issue.

In regard to Tillandsia ant-plants, Dr Benzing notes that their dry windblown seed offer no inducement for ants to disperse or collect seed for ant gardens, which is perhaps why some members of the so-called pseudobulbous clade have developed ant-housing rather than a life in ant-gardens. Furthermore, because trichome coated Tillandsia leaves are better than their roots in regard to nutrient uptake this also predisposes these plants to ant-fed mutualisms. He also writes that “ant-fed, ant-house Bromeliaceae outnumber those that root in ant-cartons.” Yet there is little or no evidence that ‘ant-house’ bromeliads can “equal other ant-house flora.” Presumably this is referring to their relative lack of evolutionary development toward symbiotic mutualisms. Nonetheless, as we will see in my next article on ant-gardens, probably more bromeliads gain reproductive vigour from a generalised hosting of ants and their wastes than does any other plant family.

Incidentally, it was Dr Benzing that as long ago as 1970 first confirmed the absorption of ant derived nutrients within Tillandsia domatia.

Tillandsia Gleanings.

T. atroviridipetala is a tiny species that has a true bulb not a pseudobulb so it does not house ants.
T. baileyi are considered to be facultative ant-house plants with a field recorded occupancy rate of 30%.
T. balbisiana are ant-house plants (occupancy rate of 42%) and in south Florida they provide nectar for ants on the primary bracts of their immature inflorescences, (Koptar 1992.)
T. bulbosa is an ant-house species (41% inhabited) that exhibits unexpected tolerances for shade and humidity and being able to absorb CO 2 as readily if wet overnight as when dry. This is probably because it lacks the layer of absorbent trichomes on exposed leaf surfaces which are typical for most other air plant Tillandsia that cannot withstand deep shade or abundant moisture, yet bromeliads as a whole are quite versatile.
T. butzii is another ant-house species that has the same flat, rigid, trichome shields as T. bulbosa which helps both of these species grow and reproduce in deep shade. Their trichomes pass rather than scatter photons whether wet or dry, they also shed rather than absorb water, permitting better survival in consistently damper habitats, which is a problem for most xerophytic Tillandsia which can suffocate if their trichomes remain too wet for too long.
T. caput-medusae is an ant-house species that grows in some of the same regions as the ant-house orchid Myrmecophila christinae var christinae (syn. Schomburgkia tibicinis.)
T. disticha is an ant-house plant according to Rauh. 1970.
T. ehlersiana is also an ant-house plant according to Rauh 1970.
T. filifolia is not an ant-house plant but a small species with very thin leaves and a true bulb.
T. flexuosa, T. paucifolia, and T. utriculata are epiphytic in Florida and parts of Mexico & Central America where they have exclusively sexual inflorescences. However, T. flexuosa occurs in coastal Venezuela where it grows terrestrially. There it is viviparous, meaning its seed will germinate while still on a parent plant.
T. fuchsii (synonym T. argentea) is another tiny species with a true bulb.
T. ionantha lost more than half of its water reserves but was able to commence photosynthesis very shortly after watering.
T. paleacea, T. purpurea and T. werdermannii hail from the extremely arid deserts of coastal Peru where survival outside of rare La Nina wet years is dependent upon the region’s regular fogs and dews. Therefore, they may benefit from CAM driven increases in early morning osmotic potentials.
T. paucifolia. Ant-house species such as this growing in dry Honduran forests would receive half or more of their entire annual (non ant) nutrient intakes in only one to eleven rainy days in each year. Ant occupancy rates of only 10- 15% have been recorded in southern Florida.
T. plumosa has a true bulb so is probably not an ant-house plant.
T. pruinosa is a densely-trichomed pseudobulbous species. In its Florida, USA habitats, its habitat niches are in shaded understory sites free of competing Tillandsia species that almost invariably prefer higher, brighter positions. Perhaps this is why it has not been confirmed as an ant-house plant. Ants also prefer warmer sunnier sites.
T. pseudobaileyi is an ant-house plant.
T. recurvata is recorded as losing 64% of its water stores but was able to regain full photosynthetic activity soon after watering.
T. seleriana is a particularly robustly-based ant-house species.
T. streptophylla is a night flowering ant-house species with an ant occupancy rate of 53% of plants checked.
Along with T. balbisiana and T. bulbosa it also provides excellent protection of home trees from a very damaging beetle. Indeed, protection from insect browsers is so efficient that researcher Alain Dejean (1992 cited by Benzing 1990.) suggested that bromeliads containing ant colonies be placed in orchard trees.
T. tectorum is a particularly beautiful, white-trichomed species with clumps of long thin leaves that grows high in the Peruvian Alps where locals often grow it on their roofs it is so attractive. The regions aridity ensures that this species must be extremely water use efficient, yet it has thin leaves, not at all succulent hence has very little water storage abilities. Therefore, it is hypothesized that its leaf structure and trichomes are best designed to make highly efficient use of quickly-passing fogs. It is easily over-misted in cultivation.


Benzing, D. 1990. Vascular Epiphytes: General Biology and Related Biota. (Cambridge Tropical Biology Series.) Cambridge University Press.
Benzing, D. 2000. Bromeliaceae: Profile of an Adaptive Radiation. Cambridge University Press.
Bluthgen, N. Verhaagh, M. Goitia, W. Bluthgen, N. 2000. Ant nests in tank bromeliads – an example of non-specific interaction. Insectes soc. 47 pp313-316.
Chew, T. De Luna, E. Gonzales, D. 2010. Phylogenetic Relationships of the Pseudobulbous Tillandsia species (Bromeliaceae) Inferred from Cladistic Analyses of ITS 2, 5.8S Ribosomal RNA Gene, and ETS Sequences. Systematic Botany (2010). 35 (1) pp 86- 95.
Rauh, W. 1970. Bromeliads for Home, Garden, and Greenhouse, Poole Blandford Press, pp. 28-31.
Rowe, D. J. 2010. Ant-plants: Arboreal Wonders of Nature. A book distributed only as DVD. See;
An interesting YouTube introduction to the Venezuelan lost worlds can be seen at;

Updated 21/09/11