Exormotheca bulbigena sp . nov . ( Hepaticae , Marchantiales ) and its relation to E . holstii in southern Africa

A new species Exormotheca bulbigena is described from southern Africa and its relation to E. holstii Steph. is discussed. Morphologically these species arc very similar and can be distinguished only when fertile. The chromosome numbers, however, n = 32 for E. bulbigena and n = 18 for E. holstii, distinguish sterile living plants.


sank E. megastomata under E. holstii.
Chromosome numbers of Exormotheca samples from the loci classici of Marquand and of Amell and from other sites in eastern Africa as well, always yielded a chromo some number of n = 18. However, Exormotheca plants collected by Volk at Gaikos and Otjua in Namibia have a chromosome number of n = 32. Nevertheless, all other morphological characters of the sterile plants were iden tical to those from the eastern sites. A priori, different chromosome counts cannot be considered as distinctive on species level, because there are many different species in the Marchantiales with the same chromosome number. On the other hand, some species with as many as six different caryotypes are also reported (Bomefeld 1989;Fritsch 1982).
By cultivation in a greenhouse, Volk succeeded in growing fertile plants from material collected in the east ern localities and in Namibia. The present study is based on the examination of living and fertile plants, and only those specimens are considered for which the chromo some number and/or the spores are known. The study of the sexual organs and their products and the asexual re production of Exormotheca has shown that the sinking of E. megastomata under E. holstii (Perold 1994 However, an additional new species, E. bulbigena, has to be established.

MATERIALS AND METHODS
Dry herbarium samples of Exormotheca from southern Africa were cultivated in a greenhouse on a mixture of garden mould and sand over a base of peat.
For chromosome counts, thallus tips were fixed, extracted and stained with orceine as described earlier (Bomefeld 1984). In the present study only samples with known chro mosome numbers are considered, although the number of Exormotheca localities is far greater (Perold 1994).
The localities are listed according to Edwards & Leistner (1971).

DISCUSSION
In her study of Exormotheca holstii samples from southern Africa, Perold (1994) necessarily used dry ma terial from various localities collected in different years. In our studies, samples of E. holstii from Namibia and Mpumalanga (eastern Transvaal) and of E. bulbigena, grown under identical conditions, developed carpocephala ( Figure 1A, B). Other characters such as overall size and shape, shape in cross-section, cell shape of the assimilatory tissue, width and height of the stomata [the papillae on top of (or) side of which stomata are located], shape of the ventral scales, and colour were all the same. Be cause these characters are meticulously described by Perold (1994) they are not repeated here, where the dif ferences between the two species are emphasized. The only morphological feature which differs in the sterile thalli are the pores of the stomata; they are more elongate in E. bulbigena than in E. holstii ( Figure 1C, D; Table 1). Whether or not these relations are affected by ecological factors in the field is unknown and therefore these differ ences should not be overestimated.
The formation of small bulbs is known for E. tuberifera (Kashyap 1914) and for Corbierella ( = Exormotheca) algeriensis Douin & Trabut (1919). Therefore we investi gated our cultures for such organs and indeed found these in E. bulbigena (inde nomen). They are cushion-shaped, 2-4 mm long and broad, and 2-3 mm thick. The bulbs consist of a chlorophylliferous, spongy tissue with some oil cells, somewhat more compact ventrally. When mature these little bulbs are no longer attached to the thallus and occur isolated in the ground. When dry they shrivel up and are easily overlooked in the field. Figure IE & F shows the dorsal and ventral view of bulbs which were rehydrated for one day after seven months of desiccation. On the lower side the remnant of the stolon which formed the bulb is visible (arrow). Exormotheca holstii also forms short stolons with a slightly enlarged terminal bud, but these are not drought tolerant and upon remoistening be come covered by mould.
As there appears to be little variation in the caryotype, the main character for discrimination of sterile plants remains the chromosome number: n = 32 for Exormotheca bulbigena and n = 18 for E holstii (Figure 2A, C). Chromosome analy sis (Bomefeld 1984) reveals that E. bulbigena is eutetraploid to a basic number of eight chro-mosomes and E holstii is eudiploid to a basic number of nine chromosomes ( Figure  2B; D). The difference in the basic numbers is not surprising when taking into account that for Exormotheca fimbriata (Brazil) both caryotypes are reported: n = 8 (Jovet-Ast 1976) and n = 9 (Bomefeld unpubl.).  The type of Exormotheca holstii is dioecious; the types of Marquand (1930) and of Amell (1953) are sterile. In our cultures out of 12 samples from South Africa only one contained both sexes; four were female, three were male, and four remained sterile. When comparing fertile plants of E. bulbigena and E. holstii the most obvious difference is the deep sulcus caudal to the carpocephalum of the monoecious E. bulbigena ( Figure 1A) which is lacking in the female E. holstii plant ( Figure IB). Closer inspection of the sulcus shows the antheridial necks which are composed of eight rows of cells ( Figure 3A). The whole structure has a diameter of ± 160 (im. The anthendial necks of E. holstii ( Figure 3B. D) have a diameter of only 100 |im. A series of sagittal sections of the car pocephalum of E. bulbigena ( Figure 4A, B) reveals the presence of two rows of additional 'microantheridia' 40 x 30 ^im; the respective values for the main antheridia in the depth of the sulcus, which are arranged in two parallel rows, are 350 and 150 Jim. These microantheridia are a unique feature not described as yet for liverworts. The reinforcing bands in the sporangium wall are very variable within one sporangium and thus are of no value for dis tinguishing between the two species. The spores of both species are about the same size (120-150 p.m), dark brown to black, anisopolar, without a wing. The ornamentation of the distal face of E. bulbigena can be described as 'vermiculate' (sensu Perold 1989) ( Figure 5A). If the ridges become very short (e.g. Perold 1994, fig. 6A) it can form a papillate pattern. The corresponding structure of E. holstii, very broad, indented papillae, could be de scribed as 'polygonal'. Figure 5C shows a spore of the sample Crosby 1115 from Zoetvlei and its similarity to that of the type of E. hobtii ( Figure 5E) proves the cor- rectness of the identification by Perold (1994). In the de-mal face of the spores is rather fine and the triradiate mark scription of E. holstii, with respect to the papillae of the is inconspicuous ( Figure 5B, D, F). The differences of the spores, Stephani (1899) mentions 'papulis saepe rostratis'. distal face become visible in another way by dark-field In Figure 5E no rostrum-like structures can be detected microscopy where the spores are seen in a bright orange on the papillae and it remains unclear to which structures colour. With this method structures are visible on the Stephani refers; elaters stuck to the papillae probably led ridges of the ornaments which in £ bulbigena are puncto the remark mentioned. The ornamentation of the proxi-tiform or very short grooves, for E. holstii they may be described as long, finely branched furrows ( Figure 6A, B). The elaters of both species are degenerate, about 20 pm in diameter and between 50 and 150 (im long, with rings or incomplete spirals. Summarising all the differ ences between E. bulbigena to E. holstii (Table 2) we consider it obligatory to consider the former as a distinct species, even though sterile plants are very similar.
Consideration of the climate diagrams (Walter & Lieth 1967) for typical sites of the two species ( Figure 7A, B) suggests that E. bulbigena is better adapted to a hotter and drier climate than E. holstii. The site of the latter at Rietfontein in Namibia is near a fountain and thus not contrary to this suggestion.
The immediate influence of external ecological factors on physiological activity of these plants is difficult to es timate. Both species form droplets of condensed water inside the stomata which consist of living cells. By cool ing down the air within the stomata by evaporation and  by their sheltered humidity the plants seem to establish a greenhouse-like microclimate of their own, certainly fa vourable for photosynthesis. Figure 8 shows the distribu tion of the two species in southern Africa. So far K bulbigena has only been found in Namibia and may thus be considered endemic to this region.