SPECIATION

Almost nothing is known about sexuality in Hypoxylon. Only one species has formed the teleomorph in culture. However, a number of xylariaceous fungi have completed their life cycles in culture--including species of Kretzschmariella, Nemania, Poronia Willd., Rosellinia, and Xylaria--and all are homothallic. Attempts to create mutants and genetic markers and to cross isolates have failed. However, a few species are known to produce trichogynes along with abundant conidia (Rogers, 1979). It is hypothesized that ancestral Xylariaceae were heterothallic with a trichogyne-spermatium system. Over time, homothallism has become predominant if not universal; conidia have persisted, often as propagules. The colony of the general Hypoxylon is somewhat reminiscent of the strawberry-coral model of Williams (1975) in that asexual or clonal expansion is supplemented by sexual reproduction. The amount of genetic variation among ascospores of any Hypoxylon colony is, of course, unknown.

It is probable that Hypoxylon has evolved and speciated along with the angiosperms, as discussed by Rogers (1979). Speciation has probably occurred by adaptation to a given host or hosts along with reproductive isolation that occurred in a shift from heterothallism to homothallism and/or by geographic or ecological isolation of the host(s) itself. The actual relationship of most Hypoxylon species to their hosts is as a weak pathogen with high saprophytic capacity (facultative parasite) or as a virulent pathogen with low saprophytic capacity (facultative saprophyte). Some Hypoxylon species are probably endophytes, but much of what has been called endophytism in Hypoxylon sensu stricto is actually latent infection in a highly predictable host, e.g., H. fragiforme on Fagus (Chapela et al., 1990). Indeed, true endophytic proclivities might be highly useful in separating Hypoxylon and its immediate allies, Daldinia and Entonaema, from Nemania and many Xylaria species.

There is little question that certain Hypoxylon species have followed their hosts for long distances via the activities of plate tectonics and, probably, island-hopping. For example, Hypoxylon bovei was described from Nothofagus in South America. It is highly associated with Nothofagus spp. in New Zealand. Nothofagus is a well-known South American-Australasian disjunct genus (Thorne, 1972). According to Thorne (1972), about 78% of angiospermous families, 24% of genera, and 1% of species are widely disjunct. To the extent that Hypoxylon taxa are genus-associated--and this appears to be true on few data (see TAXA KNOWN OR APPEARING TO HAVE A RESTRICTED HOST RANGE)--it is expected that speciation occurred as hosts and their associated Hypoxylon taxa were moved from the leveling effects of large populations and/or to different climatic zones where favorable mutations or sexually generated variants (if any) could persist and increase (Sewall Wright phenomenon). There appear to be more Hypoxylon species in the American tropics than elsewhere. This is probably host-driven, but the extent is unknown because of the dearth of host identifications. One intriguing question is the relatively large number of Hypoxylon species reported from South America as compared with Africa. Indeed, Africa is the only continent lacking a Hypoxylon species that is known only from that particular continent (see GEOGRAPHICAL DISTRIBUTION OF CERTAIN TAXA)! It has been assumed that African and South American flora differ considerably, with that of Africa being relatively depauperate. According to Gentry (1993), South American and African floras are much more similar than usually supposed. However, during the latter half of the Cenozoic the earth's climatic deterioration, coupled with geologic uplift of Africa, led to a drier climate and major impoverishment of the African forest flora, culminating in massive extinctions during the Pleistocene dry periods and today's very different floras (Gentry, 1993). At the community levels, floras of Africa and South America are said to be much alike; however, the African communities are composed of the same species to a much greater extent than are neotropical communities (Gentry, 1993). It thus seems probable that some Hypoxylon taxa that were early introduced to Africa did not survive owing to climatic and geological changes and loss of hosts.

New Zealand has a number of Hypoxylon species not known to occur elsewhere and even representatives of taxa known elsewhere seem to differ somewhat. This is also true for Xylaria of New Zealand (Rogers and Samuels, 1986). This is unsurprising, given the unique native flora and its long-time isolation. More surprising, perhaps, is the comparative lack of Hypoxylon and other xylariaceous fungi-in terms of both species and encounters-in Sulawesi (as represented by North Sulawesi) as compared with Borneo, i.e., opposite sides of Wallace's line (Rogers et al., 1987).

Burger (1992) has commented on the altitudinal parapatry among phanerogams on Costa Rica's Caribbean slope. Congeneric species occurring together on the same slope are usually not closely related sister species (Burger, 1992). He hypothesizes that, for some species, the interaction of host, environment, and pathogens can create sharp limits to survivorship on a gradual altitudinal gradient, i.e., could be the result of temperature and moisture on host resistance, pathogen virulence, pathogen presence or absence, or the restricted range of pathogen vectors (Burger, 1992). As a corollary to Burger's hypothesis (1992), it might be that the occurrence of a pathogen at the edge of a host range could likewise contribute to endemic speciation of the pathogen, e.g., Hypoxylon. There are no data, however, to support this.

There appears to be considerable endemism in Hypoxylon and additional intensive collecting will undoubtedly reveal much more. Such endemism is undoubtedly influenced by mountain ranges and large bodies of water and host distributions, particularly in jungles and at high altitudes at lower latitudes. To what extent neotropical jungles-with their large average number of phanerogamic species per hectare-influence speciation in Hypoxylon and other fungi is unknown.

In delimiting taxa of Hypoxylon we have put considerable weight on stromatal color as well as the color of stromatal granules in KOH. The extent to which these colors become "fixed" within a putative species is unknown. Is pigmentation in Hypoxylon or any other fungus subject to change under direct competition? Schluter (1994) has shown that stickleback fish species, when coexisting, are quite distinct. When a stickleback species has a habitat to itself, it takes an intermediate form. Other examples are given by Weiner (1995).

It has been observed that, when species have variants that can be equated to varieties, the smaller-spored variety generally occurs at a lower latitude, i.e., in a warmer overall climate. This perhaps indicates that, for a given taxon, large ascospores are most advantageous under cooler climates, i.e., climates wherein ascospores might be subjected to a long dormancy or quiescence requiring respiratory substrate prior to germination. This holds only for taxon pairs. Among species, some of the largest ascospores occur in the tropics!

There are in Hypoxylon pairs or groups of taxa whose members are separated from one another on only one or several characters. For example, Hypoxylon anthochroum is a widespread species that is separated from H. placentiforme, another widespread species, primarily on the conspicuously thick stroma of the latter. It differs from H. rubiginosum and H. lividipigmentum in the color of stromatal pigments in KOH. Hypoxylon rubiginosum, in our concept, has been found mainly in North America and Europe, whereas H. lividipigmentum is known only from Central America and South America. Moreover, it is separable from H. fuscum on anamorphic characters and on the great general specificity of the latter for Alnus. It can be hypothesized that-if these taxa are truly related-they came from a genotypically and phenotypically plastic early ancestor. Certain taxa such as H. anthochroum and H. placentiforme early became isolated in some way-perhaps by the apparent reduction of outcrossing by homothallism-and became differentiated by stromatal morphology. These taxa were then distributed by continental rafting, etc., in similar ranges. Taxa such as H. rubiginosum and H. lividipigmentum might have early become geographically isolated from the common ancestor. Hypoxylon fuscum might have become isolated from the common ancestor by host specificity and largely became distributed along with its Alnus host. It is to be noted, however, that H. fuscum is a complicated species that probably will eventually be broken into additional taxa. Hypoxylon fuscum has apparent interfaces with highly similar taxa such as H. macrocarpum, H. subrutilum, and H. fuscopurpureum. A number of other similar examples could be discussed, as suggested by the SUMMARY OF TAXON PAIRS AND CLUSTERS.

A number of pairs are separable primarily on ascospore size. As discussed earlier, these phenomena might be in part related to latitude and altitude. In the case of H. fragiforme and H. howeianum, however, ascospore size difference is associated with host specificity. Although both occur on Fagus, Hypoxylon fragiforme is highly associated with it, whereas H. howeianum has a much broader host range. Not surprisingly, H. howeianum also has a much broader geographical range than H. fragiforme.

There are presently almost no molecular data bearing on speciation of Hypoxylon. However, Yoon and Glawe (1993) have shown by means of RAPD studies that H. annulatum and H. truncatum are indeed distinct taxa and not synonymous as believed by Miller (1961). Careful weighing of morphological data reinforces their view.