Yes, you thought there were only three, namely Pitcairnioideae, Tillandsioideae, and Bromelioideae, known by their winged seeds, plumose seeds, and fleshy fruits, respectively. But for the last 10 years much research has been conducted to test whether this classification reflects actual evolutionary history, as reconstructed based on variation in the DNA carried by their chloroplasts.
Things are now taking shape and you should be aware of where things are heading. Thomas Givnish and his team at the University of Wisconsin recently published a paper in 2007 in Aliso, where the proceedings of the Third International Congress on Monocot Evolution (held in California nearly five years ago) have appeared in two special issues.
Remember we are considering how bromeliads evolved millions of years ago, when North and South America were separated, when the Venezuelan highlands were lowlands ready to be pushed up, and when the Amazon drained through Lake Maracaibo!
Givnish and his colleagues sequenced ndhF, a rapidly evolving gene found in the chloroplast, in 35 bromeliads and 16 closely related monocots to infer relationships among present-day bromeliad genera. They found that the long-recognized subfamilies Tillandsioideae and Bromelioideae were each monophyletic - that is, each subfamily included all the descendents of a single ancestor. Surprisingly, however, Givnish et al. showed that the subfamily Pitcairnioideae was strikingly paraphyletic, with both tillandsioids and bromelioids arising from within it. Given the ladder-like family tree the investigators recovered, they found it necessary to recognize EIGHT subfamilies, if each were to be properly monophyletic and easily diagnosed in terms of recognizable morphological characters.
As a result, they described four new subfamilies, recircumscribed Pitcairnioideae and Navioideae, sunk Ayensua into Brocchinia, and described a new genus Sequencia (named after it having been recognized initially based on its DNA sequence). The subfamily Brocchinioideae is basal-most, sister to all other subfamilies, followed by Lindmanioideae; both of these groups are restricted to the ancient Guayana Shield of northern South America. Above these subfamilies is an unresolved, three-way branch involving Hechtioideae (from Central America), Tillandsioideae, and the remaining bromeliads (involving subfamilies Navioideae, Pitcairnioideae, Puyoideae, and Bromelioideae, in their respective order of branching).
Based on the extent of genetic divergence found among present-day bromeliads, calibrated against the amount of such divergence among various groups of monocots, Givnish and his colleagues inferred that bromeliads arose roughly 70 million years ago, as terrestrial plants with C3 photosynthesis, on moist infertile sites in the Guayana Shield. Subsequently, they spread centifugally in the New World, and reached tropical West Africa (in the form of Pitcairnia feliciana) via long-distance seed dispersal some 10 million years ago.
Modern genera and subfamilies began to diverge from each other 19 million years ago, implying a great deal of evolution (and most likely, a lot of extinction) during the 51 million years of time since the ancestor of all bromeliads (and only bromeliads) arose 70 million years ago. Bromeliads appear to have begun invading drier areas in Central and South America beginning roughly 15 million years ago, at the same time as bromeliads underwent a major adaptive radiation involving the repeated evolution of epiphytism, CAM photosynthesis, impounding leaves, several features of leaf and trichome anatomy, and an accelerated rate at which new genera subsequently appeared. Givnish and his team call this the "bromeliad revolution", and it appears to have occurred just after the uplift of the northern Andes and the shift of the Amazon to its present course. They suggest that epiphytism may have accelerated speciation by increasing the ability of bromeliads to colonize along the length of the Andes, allowing bromeliads to occupy a cloud-forest landscape punctuated frequently by drier valleys. Avian pollination (mainly by hummingbirds) appears to have arisen at least twice about 13 million years ago, at about the time hummingbirds themselves were diversifying; insect-pollinated, relatively small flowers (like those in Brocchinia or Lindmania) were ancestral. Despite their representing three different lineages, members of Hechtia, of Puya, and of Abromeitiella-Deuterocohnia-Dyckia-Encholirium have evolved a suite of several different leaf and trichome traits in parallel, apparently as convergent adaptations to drought.
The new subfamilies with their genera are as follows
Brocchinioideae Givnish, subfam. nov.-TYPE: Brocchinia J. H. Schultes.
Included genus: Brocchinia
Lindmanioideae Givnish, subfam. nov.-TYPE: Lindmania Mez.
Included genera: Connellia, Lindmania
Hechtioideae Givnish, subfam. nov.-TYPE: Hechtia Klotzsch.
Included genus: Hechtia
Puyoideae Givnish, subfam. nov.-TYPE: Puya Molina.
Included genus: Puya
Navioideae, descr. emend.
Included genera: Brewcaria, Cottendorfia, Navia, Sequencia, Steyerbromelia
Pitcairnioideae, descr. emend.
Included genera: Abromeitiella, Deuterocohnia, Dyckia, Encholirium, Fosterella, Pitcairnia
KEY TO BROMELIAD SUBFAMILIES
| l. | Fruits indehiscent, baccate | Bromelioideae |
| Fruits dehiscent, capsular | 2 | |
| 2. | Seeds plumose-appendaged | Tillandsioideae |
| Seeds winged or naked | 3 | |
| 3. | Flowers dioecious, plants of Central America | Hechtioideae |
| Flowers perfect, or rarely monoecious or polygamodioecious, or dioecious and plants of the Brazilian Shield | 4 | |
| 4. | Petal blades showy, tightly spiralled after anthesis, broad and distinct from claws | Puyoideae |
| Petal blades remaining free after anthesis, or if slightly coiled, then not clawed | 5 | |
| 5. | Petals large and conspicuous or, if minute, then sepals imbricate and anthers basifixed, linear | Pitcairnioideae |
| Petals minute and sepals cochlear, or petals and bracts various and sepals convolute | 6 | |
| 6. | Sepals convolute | Lindmanioideae |
| Sepals cochlear and petals minute | 7 | |
| 7. | Leaves entire, stellate chlorenchyma abundant | Brocchinioideae |
| Leaves toothed, stellate chlorenchyma absent | Navioideae |
Am I being premature in bringing these proposed changes to the notice of the layman? I think not, even though the study of DNA is in its infancy it is revealing some interesting results that we should all be aware of. I have used these findings to create a new Key to the Bromeliaceae genera which these days seems to be a yearly chore because of changes. In fact it is a never-ending story. Currently, an international consortium involving labs in the US, Austria, Germany, England, Panama, and Australia are compiling data on several different chloroplast genes and nearly 100 bromeliads to test the new classification.
Acknowledgments
The author would like to thank Tom Givnish in helping him try to understand the technical parts and present it in a more readable form!
Literature cited
Givnich, T. J,, Millam, K.C., Berry, P.E., and Sytsma, K.J. 2007, Phylogeny, adaptive radiation, and historical biogeography of Bromeliaceae inferred from ndhF sequence data Aliso 23, pp. 3-26, Rancho Santa Ana Botanic Garden>