Pantoporate pollen in the Asteraceae (Vernonieae)

Abstract Pantoporate pollen, which occurs sporadically in the Monocots and Dicots, has now been found in Asteraceae in two apparently related genera of the tribe Vernonieae, Polydora Fenzl and Oocephala H.Rob. Disposition of pores in Polydora seems more asymmetric than in Oocephala. Based on the known relationships within the Vernonieae, some conjectures are made regarding development of the pantoporate condition in the group.


Introduction
Th e tribe Vernonieae is notable for remarkable variations in pollen structure, variations in sublophate and lophate forms with varying degrees of reduced perforated tectum in the exine and variations in the attachments of the outer exine to the footlayer. Now a new variation has been encountered during a study of South African Vernonieae that involves the fi rst known examples of pantoporate pollen with non-equatorial pores in the Asteraceae. It is of some interest that one of the genera involved, Polydora (as Crystallopollen) was one in which pollen was fi rst used as a taxonomic character in the Asteraceae by Steetz (1864). Th e new variants are occasion for some discussion of the positioning of pores in the pollen of Angiosperms.
Already known in the Tribe Vernonieae are the only examples of pollen in the Asteraceae with perforated tectum partially or completely lacking in non-colpar areas. Also unique to the tribe in the family is the six-porate form of pollen in the southeast Asian genus Camchaya Gagnep., where the pores occur equatorially in three pairs (Bunwong andChantaranothai 2008, Robinson andSkvarla 2010). Th ere is also a great deal of variation in lophate patterns in pollen of the tribe. Th e principle mechanism marking orientation of poles and pores and off ering the most obvious provision for expansion of the grains when wet, the colpi, are lost in a set of exclusively paleotropical subtribes. As a result, many of the genera of the Vernonieae can be distinguished by the form of their pollen (Robinson and Skvarla 2010). Now, for the fi rst time in the Vernonieae or Asteraceae we report a further variation involving pollen with pores that are pantoporate (global distribution), that is, not all equatorial in position.

Materials and methods
Specimens examined are from the U.S. National Herbarium in Washington, D.C. Examination of pollen with a light microscope was insuffi cient to show the position of the pores. Pollen grains in most illustrations were treated with acetolysis (Erdtman 1960), followed by staining with osmium thiocarbohydrazide solutions and sputter coating with gold-palladium (Robinson and Skvarla 2006, 2007, Robinson et al. 2008. Unacetolyzed grains were rehydrated in water or alcohol directly from herbarium sheets and similarly sputter-coated. Observations were made with a JEOL 880 (Samuel Roberts Microscopy Laboratory at the University of Oklahoma) scanning electron microscope (SEM) or a Leica 440 (United States National Museum of Natural History in Washington scanning electron microscope, both equipped with lanthanum hexaboride (LaB 6 ) electron sources. To achieve limited increased breakage, material gathered at the base of a centrifuge tube was given a few jabs with a dissecting needle.

Results
Th e two genera in which the pantoporate pollen is now being reported are both restricted to sub-Saharan Africa. Th e two genera, Oocephala and Polydora, are evidently closely related to each other according to the pollen similarity, and both have pedunculate heads, a characteristic of many genera. Nevertheless, the two genera are easily distinguished from each other. Polydora has an ordinary broadly campanulate capitulum, and an ordinary capillary pappus. Oocephala (egghead) has an egg-shaped capitulum and a plumose pappus.
Examined material (Table 1)  Th e pollen of both genera has a lophate exine pattern that is nearly spherically symmetrical. Th e pattern consists of a mixture of pentagonal and hexagonal lacunae. Perforated tectum is essentially absent leaving totally exposed muri and their subtending columellae. Both genera also share a mural structure that has been referred to as rhizomate (Robinson 1999) (Figs 1F, 2F, 3C, 4C, 5B, D, 6C, E). In these muri, the columellae, instead of being fi rmly attached to the footlayer at their bases, are attached basally to each other by a lower horizontal structure (rhizome) that is itself weakly attached to the footlayer. Such a structure is not unique to the two genera studied, but occurs in many lophate, triporate, non-colpate pollen grains in the subtribes Centrapalinae and Erlangeinae that are native to the Paleotropics. A note of interest is the possible diff erence in the chemical make-up of these muri, since they have been seen to survive the degrading eff ects of chloral hydrate more completely than the muri and tectum of other Vernonieae when mounted in Hoyer's solution (Anderson 1954).
Positions of the pores are often diffi cult to determine in lophate pollen of the Centrapalinae and Erlangeinae. Th e lacunae in which the pores occur are totally undifferentiated in any other way, being either pentagonal or hexagonal, even in the polar        regions. Th e depths of the lacunae usually prevent seeing more than one pore at a time unless the grains are unacetylized and the caps on the pores are preserved ( Fig. 2A) . Th us, the easily dehiscent muri of Oocephala and Polydora are particularly helpful. Th e totally exposed sphere of the pollen footlayer shows the pores very well (Figs 1C,D,2E,3C,4C,6C,7C,D). In this way, positions and numbers of pores on the grains can be more accurately seen (Figs 1C, D, 2E, 7C,7D). To help in the estimates, a toy ball was obtained from the museum shops that showed a pattern of pentagons and hexagons that was a very close approximation of the pattern of lacunae in the pollen grains of Oocephala and Polydora. It was study of the toy ball that led to the conclusion that the pollen of Oocephala could characteristically have seven or eight pores. (Figs 1C, 2E).
An exact count of the pores can never be certain, since there is evidence of some uneven distribution of pores in some grains. In a few cases, in Oocephala, even in grains with the muri present, pores can be seen in adjacent lacunae (Figs 2A, B, 3A, 4B, 6D). In other cases in Polydora (Figs 3C-7C), expanses can be seen that have no pores (Fig. 5B, C, E). Th e total number of pores may never be more than the fi ve or six that can be seen in some views of Oocephala (Figs 1C, D, 2E), and asymmetry is probably even more extreme in Polydora (Fig. 5C, E), but pantoporate non-equatorial distribution of pores is certain in both genera, and at least some asymmetry is certain.

Discussion
Two basic porate conditions of Angiosperm pollen are well known (Walker 1974). Th e forms derived generally from the tetragonal or "cruciate" division of the pollen mother cell that is found in Gymnosperms, Monocots and basal Dicots. Th e alternative tetrahedral or decussate division of the pollen mother cell is basic to the Eudicots, a sort of reversion to the spore mother-cell pattern of division found in most Cryptogams. Th e polyporate (pantoporate) condition is found sporadically in all three Angiosperm groups, or as stated by Harley (2004) "known from all three major groups of Angiosperms, although infrequently in the monocots and even less frequently in the basal dicots". In monocots (Harley 2004, Nadot et al. 2006 It is reported in some Bromeliaceae, Alismataceae and Araceae. In basal dicots polyporate pollen is mentioned only from the Trimeniaceae (Sampson 2000, Harley 2004). In the eudicots , Harley mentions Caryophyllales, Ranunculales, Podostemaceae, Linaceae, Zygophyllaceae and Geraniaceae as achieving the polyporate condition by multiplication of simple furrows that are reduced to pores. In other families such as Malvaceae, Sterculiaceae, Th ymelaeaceae, Euphorbiaceae and Malpighiaceae a spiralization of a line joining centers of apertures is suggested. All of these are in derived groups, and polyporate pollen is not basic to any group.
Th ere have been no previous reports of pantoporate pollen in the Asteraceae, and suggestions here are based on observations of the pollen of related members of the family in the tribe Vernonieae.
Th e Asteraceae has pollen that is basically tricolporate, a condition that seems basic to all the tribes including the Vernonieae (Blackmore et al. 2009). Th e most important evidence comes from the fact that the polyporate condition in the Asteraceae is found in the tribe Vernonieae which has the some of the most destabilized pollen formation of any tribe in the family. Th e pantoporate pollen has developed within the subgroup of the Vernonieae that already has the most spherically symmetrical forms of pollen in the tribe, the subtribal pair, Erlangeinae and Centrapalinae. It is in these subtribes that the colpi are totally suppressed, and orderly arrangement of the lacunae, even at the poles, is completely lost. Th ese features that have become so spherically symmetrical in these subtribes are among the last layers laid down as the pollen grain matures. Th e position of the pores is much more basic, being established in some of the innermost and earliest laid down layers of the pollen. It is the contention here that the loss of radial symmetry in the Vernonieae pollen started with the outermost and most belatedly deposited layers of the grains, and progressed eventually to the inner and earlier deposited layers.
Th ere is even a developmental basis for the asymmetry in the distribution of the pores, more obvious in Polydora and less evident in the more specialized Oocephala. Th is could trace to the fact that the developing pollen grain has a distal surface that faces outward in the pollen mother cell, and a proximal surface that faces inward toward the center of the tetrad of pollen grains. Th us, the distribution of the pores is infl uenced by early stages in pollen development, beginning with position in the tetrads or mother cell.