Proanthocyanidins are rarely found in leaves or stems of members of some centrospermous families (Bate-Smith 1962; Gibbs 1974), whereas they seem to be commonly present in the seedcoat of most Centrospermae (Bittrich and Amaral 1991).
Bound ferulic acid in unlignified cellwalls is an important chemical marker for the whole order. Within dicotyledons, only members of the Centrospermae show this feature, as do a number of monocotyle-donous families (Harris and Hartley 1980; Hartley and Harris 1981).
3. Macromolecular Evidence. In a serological study, Jensen (1965) confirmed the placement of the Di-diereaceae in the Centrospermae, where they may be closely related to the Cactaceae and Portulacaceae. DNA-RNA hybridization experiments of Chang and Mabry (1974) led to the conclusion that the betalain families are more closely related to each other than to the Caryophyllaceae. Palmer et al. (1988) investigated cpDNA rearrangements of several plant families and found the loss of the rpl2 intron a useful marker for the Centrospermae. Their study failed, however, to find a conclusive hint to next-related families within angiosperms. The members of the Polygonaceae investigated lacked the marker.
In the phylogenetic trees of Rettig et al. (1992) based on rbcl gene sequences, the Amaranthaceae/Cheno-podiaceae branch off at the base of the centrospermous clade; the Molluginaceae and Caryophyllaceae form distant branches. As to the evolution of flower pigments, the authors conclude that betalains have been lost and replaced by anthocyanins secondarily. Faced with such surprising results, it seems prudent to wait for further studies utilizing molecular characters.
4. Ultrastructure. While the characteristic peripheral ring of proteinaceous filaments in the sieve-element plastids was first detected with light microscopy by Esau (1934) in Beta vulgaris, Behnke (e.g., 1976, 1981) conducted extensive TEM studies of this character. The sieve-element plastids of the Centrospermae are unique among angiosperms in possessing a subperipheral dense ring of protein filaments, only rarely containing starch inclusions (e.g., in Phytolacca, Bougainvillea, Iresine celosioides, and Macar-thuria), and showing a central protein crystalloid, which is absent only in Amaranthaceae and Cheno-podiaceae (except Sarcobatus, Behnke, pers. commun.). Outside the Centrospermae, sieve-tube plastids with proteinaceous inclusions, although of different shape, are constant attributes of the monocotyledons and of several representatives of the Magnoliales in the widest sense.
Phytoferritin inclusions in phloem tissue are another TEM character investigated by Behnke (1978) in the Centrospermae. Rodman et al. (1984) claimed the presence of these inclusions in phloem parenchyma as a possible synapomorphy supporting their clade of Aizoaceae/Cactaceae/Didiereaceae/Portula-caceae. However, as stated by Behnke (1978), these inclusions, for unknown reasons, seem to be correlated with the succulence of stems or leaves, and were found to be absent in Claytonia, the nonsucculent member of the Portulacaceae investigated.
5. Flower Characters. In the Centrospermae, different organs serve as optical attractants for pollinators. These include staminodes (many Aizoaceae, probably some Molluginaceae), tepals which are either completely petaloid (some Phytolaccaceae) or only so on the inside (often in Aizoaceae), showy bracts (some Amaranthaceae), involucral bracts (some Nyc-taginaceae), and numerous hypsophylls (Cactaceae). The morphological interpretation of the "petals" in Caryophyllaceae, Molluginaceae, and Portulacaceae is still controversial. It is noteworthy that the petaloid organs of anthocyanin and betalain families basically do not differ.
The most important flower character in the order is the ontogenetic reduction (more rarely complete absence) of the septa in the ovary and the development of a free central placenta or, if the number of ovules is low, a mostly (sub)basal placenta (Eckardt 1976). Among the families which are often assumed to be rather primitive members of the order, the Aizoaceae and Molluginaceae have completely septate ovaries, whereas some Phytolaccaceae have pseudoapocar-
pous gynoecia. The more or less unseptate ovaries of the Caryophyllaceae must be considered the result of parallel evolution. Hofmann (1977) observed differences in the ontogeny of the reduction of the septa in the Caryophyllaceae and Stegnospermaceae, which cast doubt on whether this character is homologous in both taxa. The reduction of ovule number is usually connected with a shift to (sub)basal placentation, occurring in both septate and unilocular ovaries. Exceptions can be found in the completely septate ovaries of Galenia, Plinthus, and Tetragonia (Aizoa-ceae), which have one ovule per locule with subapical placentation. On account of their pluricarpellate, and seemingly apocarpous, gynoecia, the pluricarpellate representatives of the Phytolaccaceae have often been considered as the most primitive group of the whole order. Rohweder (1965), however, came to the conclusion that the gynoecium of Phytolacca is basically syncarpous and only secondarily apocarpous in its upper part; it represents an advanced state. Hofmann (1977) supported Rohweder's view and Volgin (1988) claimed that the monocarpellate gynoecia of subfam. Rivinoideae have evolved from syncarpous ones.
Another floral character typical of centrospermous families is the centrifugal stamen inception in multi-staminate flowers, though multistaminate Mollugina-ceae and Nyctaginaceae still need to be studied in this respect. In centrifugal androecia the stamens develop either on four or five distinct fascicle primordia or on an annular meristematic swelling (ring wall). The nectary is usually located at the bases of the filaments (Zandonella 1977; Smets 1986).
Transmitting tissue for pollen tubes in the ovaries extending that of the stigmas or styles is observed in the ovaries of most families. In the case of basal placentation, the transmitting tissue may be located on a narrow septal rigde on the ovary wall (many Caryophyllaceae). It may also extend on to the funiculus (Tetragonia, Aizoaceae; Stegnosperma) and sometimes has been named funicular obturator (Prakash 1967). In the Nyctaginaceae and Rivinoideae (Phytolaccaceae s.I.), however, a transmitting tissue is located within the ovary wall, leading down to the basal ovule (Rohweder and Huber 1974; Volgin 1988). This evolutionary novelty probably is connected with the monocarpellate condition and suggests a close relationship of both taxa.
6. Fruit and Seed Characters. Various fruit types (loculicidal, septicidal, and circumscissile capsules, achenes, drupes, berries, nuts) occur in the Centrospermae, but in the presumably basal branches Caryophyllaceae and Molluginaceae (also in Aizoaceae and Portulacaceae) capsules clearly predominate. Dry indehiscent fruits mostly are correlated with a reduced ovule number. Berries or berry-like fruits, oc curring in Achatocarpaceae, Basellaceae, Cactaceae, Phytolaccaceae s.l., but only rarely in Aizoaceae, Chenopodiaceae, Amaranthaceae, and Caryophyllaceae, are probably derived. Drupaceous fruits are very rare (e.g., Tetragonia implexicoma). If the radiation of the order took place in arid habitats, certainly capsules are more likely than fleshy fruits to be the primitive type.
The sculpturing of the testa is of great taxonomic value, especially below the level of genus. The arrangement and orientation of the testa cells often follows a typical "centrospermoid pattern" (Barthlott 1984), which can be recognized in families such as the Aizoaceae, Cactaceae, Caryophyllaceae, and Portulacaceae. Little attention has been paid to the contribution by Meunier (1890) to the seed anatomy of the Centrospermae. Based on careful investigations in several families, he found various characters deserving further study; minute cutinized rodlets on the testa surface (mainly in Caryophyllaceae); so-called stalactites (cuneate stripes perpendicular to the surface) in the exotesta; and bar-like thickenings in the cell walls of the endotegmen (absent in most Caryophyllaceae and Barbeuia). Netolitzky (1926) mentioned the occurrence of a similar endotegmen in Ranunculaceae and Papaveraceae. The walls of the endotegmen in Centrospermae are often able to swell when moistened, which is important during germination. In several families, the seeds germinate by means of an operculum in the micropylar region, and the protruding radicula is protected for some time by the swollen endotegmen (Bregman and Bouman 1983; Bittrich and Ihlenfeldt 1984).
Arils of various shape and size occur in a number of families, but the distribution of arillate seeds in the order gives no clue as to whether the presence of an aril has to be regarded as ancestral or derived. The function of the arils can be very different: the arillate seeds of Moehringia (Caryophyllaceae) are adapted to myrmecochory; the bony arils of the Opuntioideae obviously have a protective function; those of Trian-thema (Aizoaceae) are sticky after wetting (myxo-spermy) and either fix the seeds to the soil or are an adaptation to epi-ornithochory.
Anomalous secondary growth is reported from a number of centrospermous families but is unknown from the Achatocarpaceae, Basellaceae, Cactaceae, Didiereaceae, Halophytaceae, Portulacaceae, and a few genera of other families. Its absence is not necessarily correlated with an annual life form. It may occur in annuals or maybe absent in perennials. Anomalous growth is reported from members of all those families which are generally considered to represent rather basal branches within the order (Aizoaceae, Caryo-phyllaceae, Molluginaceae, Phytolaccaceae). This suggests that its absence might be the derived condition in Centrospermae. Gibson and Nobel (1986) suggested that the absence of anomalous secondary growth might be a common derived character of Basellaceae, Cactaceae, Didiereaceae, and Portulacaceae. Anomalous secondary thickening is not restricted to the stem. Occasionally, it develops only in the root, while secondary growth in the stem is normal. Joshi (1937) observed that in seedlings anomalous growth appears earlier in the root than in the stem.
Was this article helpful?