Fruit type and seed morphology differentiate Bromeliaceae into three subfamilies (sensu Smith and Downs 1974, 1977, 1979; Fig. 3.6), but not as de nitively as some taxonomic descriptions imply. Dry capsules and naked or double-coated seeds with or without appendages characterize Pitcairnioideae. Seeds equipped with an elaborate ight apparatus born in capsules indicate Tillandsioideae (Figs. 3.3H, 3.6J; Chapter 12), while the berries produced by most Bromelioideae contain naked seeds equipped with or lacking soft, unbranched appendages (Figs. 3.5G,H, 3.6L). Exceptions include the fruits of Fascicularia, Ochagavia and Orthophytum (Bromelioideae), which tend toward dryness, and those of some Pepinia (Pitcairnioideae) that are just as unexpectedly eshy. Dehiscence varies among the capsular types, and at least one Ronnbergia species forcibly ejects its ripe seeds. Epigeny sometimes prevails where reports indicate hypogeny (e.g., many Tillandsioideae) and vice versa (e.g., some Pitcairnioideae).
Bromeliad seeds range from medium to small by angiosperm standards (e.g., <0.1 mg for some Pitcairnia), but none approach the proportions of the minute diaspores produced by the orchids and holoparasites. Those of terrestrial Bromelioideae exceed the sizes of the seeds of the related epiphytes if the pattern noted in similarly eshy-fruited Araceae and Cactaceae also prevails in this subfamily (Madison 1977). McWilliams (1974) determined that the seeds of Tillandsioideae generally weigh less than those of Pitcairnioideae.
Dispersal modes probably vary more among the bromelioids than among members of the other two subfamilies (Chapter 6). Mass also varies more among Bromelioideae. Seeds of some taxa (e.g., Acanthostachys, certain Bromelia, and Cryptanthus) exceed all others in size, and accordingly, ripen in smaller numbers. A pliable, sticky appendage probably effects adhesion to substrates and perhaps also to dispersers (Fig. 3.6L).
Embryos usually occupy about one-quarter to one-third of the seed
volume, with starchy endosperm (and some oil in certain taxa) making up the balance (Fig. 3.7). Development is helobial according to Davis (1966). Billings (1904) described embryology in Tillandsia usneoides as conventional for monocots, but polyembryony occurs in some close relatives (Figs. 3.6K, 3.7). Gross (1985) surveyed 11 species of Tillandsia subgenus Diaphoranthema and discovered one to four embryos in at least the occasional seed of all but T. recurvata. If more than one progeny was present, the largest of the group appeared to be zygotic and the others of undetermined origin and positioned lateral to it. More endosperm remained in seeds bearing one compared with multiple embryos.
The outermost layer of the endosperm consists of starch-free, cubical cells containing darkly pigmented, granular materials. Szidat (1922) suggested its identity as an aleurone layer. If so, component proteins, like those of the cereals, probably promote germination by mobilizing food reserves for growing embryos. Thin-walled endosperm tissue deeper in the seed contains abundant starch, usually as lenticular grains. Elongated cotyledons equipped for absorption occupy the distal end of the seed where they remain, rendering germination hypogeal (Figs. 3.7,3.8). Intercalary growth
near the base of the hypocotyl pushes part of that organ and the adjacent radicle through the testa. Seedlings of Tillandsioideae fail to produce roots for weeks to months (Fig. 3.8C). The greatest delays characterize neotenic Tillandsia.
Several features describe the ovules and seeds of Bromeliaceae, for example anatropous morphology, two layers of cells comprising each of the two parts of the integument, predominantly starch reserves, and a rel-
atively small embryo (Billings 1904; Fig. 3.7). Some endosperm always remains to nourish the young seedling. Mature seeds provide numerous potentially informative, but little-utilized, traits for taxonomy (Gross 1993). Seed morphology varies far more than most of the literature suggests.
A closer look at the development of the outer seed coat seems advisable to evaluate several suggestive similarities, including possible homologies between the ight apparatus of Brocchinia tatei and Tillandsioideae, especially Glomeropitcairnia (Varadarajan and Gilmartin 19886; Figs. 3.9, 6.1D; Chapter 12). Navia seeds exhibit an interesting parallel with those of Bromelioideae: both lose the outer integument during development, although conditions differ at maturity as described below (Fig. 3.9). The sticky strand of material (funiculus) that helps fasten the seeds of many Aechmea species (e.g., A. angustifolia; Fig. 3.6L) to substrates appears to be derived from the testa.
Searches for ever ner structure for systematic and functional analysis continue. Several classical papers (e.g., Poisson 1877; Szidat 1922; Netolitzky 1926) report modi cations of the outer testa that in uence seed mobility. However, neither these studies nor the others published since exhaust the possibilities for major revelations about family history, and how aspects of dispersal and seedling establishment favor success on spe-ci c kinds of substrates. For example, Palaci (1997) discovered that the coma of Catopsis (Fig. 3.3H) is not homologous with the ight apparatus of the other Tillandsioideae, further underscoring the isolation of this genus within its subfamily. Conceivably, Catopsis evolved capacity to disperse among aerial substrates independently. If so, epiphytism and litho-phytism, although nearly universal through Tillandsioideae, could well be homoplasious.
Was this article helpful?