| Constancea 83, 2002 University and Jepson Herbaria P.C. Silva Festschrift |
|---|
The genus Tapeinodasya, originally described by Weber-van Bosse (1904) from Indonesia and the Sulu Archipelago (Philippines), has remained one of the least known genera of the Dasyaceae. Based on recent collections from the East African coast, vegetative and tetrasporangial structures are described in detail for the first time and the placement of Tapeinodasya in the Dasyaceae based on cellulosympodial growth is confirmed. The genus is separated from all other genera of the Dasyaceae by a distinctive cauliflower-like habit, complete absence of monosiphonous filaments, extensively corticated tetrasporangial stichidia and the presence of both pre- and postsporangial cover cells. Tapeinodasya appears to be related to the Heterosiphonia-like genera, based on an alternate sequence of periaxial cell formation, a primary alternate-distichous arrangement of holoblastic branches, and branches separated from one another by one or more unbranched segments. Placement of Tapeinodasya in a clade with Thuretia and Dictyurus as recently hypothesized is verified by the recognition of a clockwise mode of alternate periaxial cell formation and a primary bilateral organisation. Comparison of the East African specimens with the type collections of the other known species, shows that the Tanzanian specimens belong to the Indonesian Tapeinodasya bornetii rather than to T. etheliae which was described from the Amirante Islands. The presence of T. bornetii in Tanzania represents a substantial range extension and a new record for the Indian Ocean.
Several studies have focussed on phylogenetic relationships of the entire family Dasyaceae, with respect to the Ceramiaceae, Delesseriaceae, and Rhodomelaceae (Falkenberg 1901; Kylin 1930; Hommersand 1963; Saunders et al. 1996; Jong et al. 1998) or on evolutionary affinities of the genera of Dasyaceae (Parsons 1975; Jong 1997). From the latter studies, however, it became evident that our knowledge of the genera of the Dasyaceae is not evenly distributed. Data on some genera are primarily based on old accounts and are incomplete in several ways, hence precluding a better understanding of the intra- and interfamilial relationships of the Dasyaceae. Collections of Tapeinodasya bornetii, the type of the genus, from the East African coast enable us to re-examine the morphology and structure of Tapeinodasya and reconsider its position in the Dasyaceae. Additionally, the specimens from East Africa were compared with the type material of T. bornetii and T. etheliae.
Branching. Three types of axes can be distinguished: 1. Bilaterally organised, compressed sympodial axes, which grow in a more or less horizontal position (Figs. 3, 4, 13); 2. adventitious sympodial axes arising on the most basal axial cell of a pseudolateral of a bilaterally organised axis; 3. radially organised indeterminate axes taking the place of a pseudolateral on radially organised axes. Although complicated at first sight and without an apparent structure, the branching system generally follows a rather strict pattern. This, however, is only clearly observed in the apical parts of the thallus. The morphology of the horizontally orientated holoblastic sympodial branches is bilateral with pseudolaterals deflected to two opposite sides of the sympodial main axis (Figs. 4, 12). The distance between two successive pseudolaterals is 375-750 µm. The axial cells of basal segments of pseudolaterals can give rise to new, apparently radially organised, adventitious sympodial axes, directed upward and of limited growth (600-1300 µm), with short distances between the pseudolaterals (100-150 µm). One of the pseudolaterals of such an adventitious sympodial axis might grow into another horizontally orientated bilateral sympodial main axis. Radially organised indeterminate axes seem to be formed at random, but are always replacing a pseudolateral of an adventitious sympodial axis. Anastomosis between axes in the lower part of the thallus is observed occasionally, but does not seem to follow a strict pattern (Fig. 9). Neither does it involve the formation of specialised connective cells. At times indeterminate axes bend down, stop forming side branches and grow towards the substrate, eventually continuing growth as prostrate axes (Fig. 8).
Apical organisation and periaxial cell formation. Except for the vertically arranged adventitious sympodial axes, which arise endogenously on the pseudolaterals, branching is essentially holoblastic. Pseudolaterals are formed every other segment, with occasionally two, rarely three segments separating successive pseudolaterals (Figs. 5, 12). Pseudolaterals are primarily deflected in a bilateral manner. In radially arranged axes however, pseudolaterals are secondarily displaced due to asymmetric cortex development. Pseudolaterals remain unbranched. Four periaxial cells are developed, cut off in an alternating clockwise sequence. After a first periaxial cell is cut off, a second periaxial cell is formed adjacent to the first periaxial cell, followed by a third periaxial cell being cut off at the opposite side of the first periaxial cell, and diametrically opposite to the second periaxial cells. A fourth periaxial cell is situated opposite to the first periaxial cell. In a next segment this order of periaxial cell formation remains exactly the same if no pseudolateral is formed. If a pseudolateral is formed the position of the first periaxial cell switches 180° (Fig. 12). The first periaxial cell is often cut off from the second segment onward; the periaxial cells are already surrounded by an extensive cortex in the subapical parts. The cortex can be up to 5 layers thick and is differentiated into a cortical and a subcortical layer in the main axes.
Reproductive structures. Tetrasporangial stichidia develop from the pseudolaterals, and consist of 16-26 fertile segments, each one producing 2 tetrasporangia (Fig. 6). Often, however, a segment may produce only one sporangium, or none at all. Stichidia are unbranched, spindle-shaped, 300-340 µm wide and 900-1700 µm long. Four periaxial cells are formed in an alternating sequence as in the axes, but only the second and third periaxial cell may cut off a tetrasporangium. A presporangial cover cell is cut off at the distal end of the periaxial cell before the latter divides transversely into an upper tetrasporangial initial and a lower stalk cell (Figs. 7, 15). Subsequently a second and a third cover cell are produced, after which all cover cells and their daughter cells divide numerous times resulting in a densely corticated stichidium, morphologically similar to sterile pseudolaterals (Fig. 16). Mature tetrasporangia are 110-135 µm in diameter.
Spermatangial branches were not observed.
Female reproductive structures. Procarps and young stages of cystocarp development were not observed. Mature cystocarps are sessile, globose (1100-1350 µm in diameter), with a short spout (Fig. 10). Carposporangia are obovate (80 µm wide, 130 µm long), maturing basipetally and formed in rows on monopodially branched gonimoblast filaments.
Tapeinodasya has traditionally been included in the Dasyaceae on the basis of the cellulosympodial growth (sensu Norris et al. 1984), the original character separating the Dasyaceae from the Ceramiaceae, Delesseriaceae and Rhodomelaceae (Falkenberg 1901; Rosenberg 1933; Kylin 1956). Unfortunately additional differentiating characters of the Dasyaceae, related to the early developmental stages of the female reproductive structures and postfertilisation events (see Rosenberg 1933; Hommersand 1963; Parsons 1975), could not be verified. However, the absence of trichoblasts and the presence of tetrasporangial stichidia are also characteristic of the Dasyaceae. Within the family the genus is easily recognised by: 1. A unique cauliflower-like habit. 2. An entirely polysiphonous and densely corticated thallus. 3. Completely polysiphonous pseudolaterals. Most genera of the Dasyaceae have pseudolaterals which are either entirely monosiphonous (e.g., Dasya, Dasysiphonia, Dictyurus, Dipterocladia, Rhodoptilum, and Thuretia) or have one to a few polysiphonous segments near the base (Dasyella, Eupogodon and Heterosiphonia). Only in some species of Heterosiphonia are pseudolaterals largely polysiphonous, with only the ultimate segments being monosiphonous e.g., H. berkeleyi Montagne, H. dubia (Suhr) Falkenberg, and H. pellucida (Harvey) Falkenberg (Ricker 1987; Stegenga et al. 1997). 4. The formation of only two tetrasporangia per segment. Although the number of periaxial cells in vegetative axes is not necessarily similar to the number of periaxial cells in stichidial axes, and partial maturing of periaxial cells in stichidia is not uncommon, in Dasyaceae the number of tetrasporangia per segment equals the number of periaxial cells in the stichidia (Jong 1997). The situation of Tapeinodasya with four periaxial cells being formed of which only two give rise to tetrasporangia is to be found only in Halodictyon (Coppejans 1975), a delicate net-forming species, of which the position in the Dasyaceae is seriously doubted (Jong et al. 1998). 5. The structure of the tetrasporangial stichidia. In the Dasyaceae tetrasporangial cover cells are cut off from the fertile periaxial cell either prior to or after the formation of a tetrasporangium. Postsporangial cover cells are found in most genera: Dasya, Dasysiphonia, Dipterocladia, Eupogodon, and some species of Heterosiphonia (Rosenberg 1933; Lee and West 1979; Kajimura 1992; Millar 1996; Jong et al. 1997; Jong et al. 1998). Presporangial cover cells are found in Dasyella, Dictyurus, Heterosiphonia, and Thuretia (Aregood and Hackett 1971; Coppejans and Boudouresque 1984; Parsons 1975; Jong et al. 1998).
The presence of both pre- and postsporangial cover cell groups has not yet been recorded in any genus of the Dasyaceae, but is fairly common in the Rhodomelaceae. Numerous members of the latter family often show two large, rectangular cover cells, produced before the tetrasporangium is cut off. A third postsporangial cover cell is usually small and triangular (Hommersand 1963). In many genera of the Rhodomelaceae, however, the additional postsporangial cover cell is absent. Because of the variation in cover cell formation, in both the Dasyaceae and Rhodomelaceae, this character is generally not regarded to be important on a family level. One more difference exists between Tapeinodasya and the other genera of the Dasyaceae with respect to the tetrasporangial stichidia. All the other genera show cover cells remaining undivided or dividing only once. In Tapeinodasya, however, the original cover cells become embedded in a 2-3 layered cortex. According to Hommersand (1963) this is an ancestral character, also present in primitive genera of the remaining families of the Ceramiales such as Carpoblepharis (Ceramiaceae) and Bostrychia (Rhodomelaceae). It therefore can be hypothesized that Tapeinodasya represents a primitive member of the Dasyaceae. Based on cladistic analysis Tapeinodasya was assigned to the Heterosiphonia-related genera (Jong et al. 1998), which are thought to contain more plesiomorphic characters of the Dasyaceae (Hommersand 1963; Jong et al. 1998).
Ancestral characters defining the Heterosiphonia-related genera present in Tapeinodasya are: a primary alternate-distichous arrangement of holoblastic branches, partly or completely polysiphonous pseudolaterals, holoblastic branches separated from one another by one or more unbranched segments, and an alternate sequence of periaxial cell formation. A phylogeny of the genera of the Dasyaceae based on morphological and reproductive characters places Tapeinodasya in an unresolved clade with Dictyurus and Thuretia (Jong et al. 1998). The latter two, and especially Dictyurus, have been considered to be isolated genera and only distantly related to Heterosiphonia. The close relationship between Tapeinodasya and the Dictyurus-Thuretia group is primarily based on the sequence of periaxial cell development and the number of periaxial cells. Additional support for the clade was given by the presence of anastomosis in all three genera. However, whereas in Dictyurus and Thuretia pseudolaterals are secondarily united to form a delicate and highly regular network, the anastomosis in Tapeinodasya, occurs in an irregular manner and never results in a lace-like network. Also the specialised lateral cells accomplishing the anastomosis of pseudolaterals in Dictyurus and Thuretia (Parsons 1975; Parsons and Womersley 1998), are absent in Tapeinodasya. The homology of this character is therefore doubted in those genera.
The Dasyaceae are characterised by holoblastic branching, in which an apical cell cuts off a subapical cell, which becomes the initial of a lateral branch prior to the formation of periaxial cells (Hommersand 1963, Parsons 1975). Due to the sympodial construction of the thallus, those holoblastic branches are modified and become sympodial monosiphonous or polysiphonous branches, called pseudolaterals. As the plane of division of the apical cell determines the side to which a branch is formed, the primary branching pattern is defined by the division of the apical cell into a new apical cell and a subapical cell. In all genera of the Dasyaceae the first periaxial cell is formed to one side of a pseudolateral, the second periaxial cell to the other side of and slightly underneath the pseudolateral (Jong et al. 1997). As a result a relationship exists between the primary branching pattern and the sequence of periaxial cell formation, limiting the number of possible patterns. On morphological grounds two different primary thallus organisations can be distinguished: a radial and an alternate-distichous pattern. Thallus architecture can be further categorised based on the morphology of the pseudolaterals, which can be entirely monosiphonous to largely polysiphonous (Fig. 22). In certain genera further modifications of these basic patterns of thallus organisation have occurred by means of anastomosis of the pseudolaterals (e.g., Dictyurus and Thuretia). In species with a primary radial structure, the position where the first periaxial cell is formed follows the spiral of pseudolaterals (often a 2/5 spiral). The periaxial cells therefore are formed in a circular sequence (Fig. 23). This is true for all primary radially organised genera of the Dasyaceae: Dasya, Eupogodon, and Rhodoptilum. Genera of the Dasyaceae characterised by a dorsiventral or a bilateral symmetry do not differ in the position of the pseudolaterals. In both cases pseudolaterals are deflected distichously. However, the sequence of periaxial cell formation differs between dorsiventral and bilaterally organised thalli. In Colacodasya, Dasysiphonia, and Heterosiphonia the first periaxial cell is always formed on the same side of the thallus, the dorsal side. The other periaxial cells are alternately formed on the left or right side of the first periaxial. The last periaxial cell is cut off opposite the first, on the ventral side of the thallus. This sequence of periaxial cell formation is referred as alternating-switching by Jong et al. (1998). In the genera characterised by a primary bilateral organisation, Dictyurus, Dipterocladia, Tapeinodasya and Thuretia, the position of the first periaxial cells shifts from the dorsal to the ventral side of the thallus, but the relative position of the periaxial cells does not switch as in dorsiventrally organised taxa. The mode of periaxial cell formation encountered in bilateral genera is referred to by Jong et al. (1998) as alternating-clockwise.