PHYLOGENETIC ANALYSIS

The object of the study was to obtain a most-parsimonious (minimized homoplasy) hypothesis of evolutionary relationships of the genera of the Pottiaceae as presently conceived, and to use this as a basis for a suprageneric classification. This evaluation is complex and a précis, as overview, follows.

SUMMARY OF THE ANALYSIS

Timmia (Timmiaceae, diplolepideous) and Polytrichum (Polytrichaceae, nematodontous) have gametophytes that are quite similar to certain genera of Pottiaceae, but are considered distant from the Pottiaceae and from each other because of their much different peristomes, which they share with other genera not in the Pottiaceae. Because these two genera are both distant and share many characteristics with the Pottiaceae, they are necessarily expected to appear at the base of any cladogram including both them and the Pottiaceae. The Polytrichales is shown to be more primitive than the Bryales in cladistic evaluations by Mishler (1986b) and Mishler and Churchill (1984) although the Buxbaumiales is the nearest sister group (but which is not used here as an outgroup because of a lack of characters that is probably due to reduction).

Of the various haplolepideous genera in families evaluated as potential outgroups (Cladograms 1–6), Ptychomitrium (Ptychomitriaceae) was the outgroup that best kept Timmia and Polytrichum low in the tree when ten non-pottiaceous genera were added to the Pottiaceae data set. Analysis was then made with only a single non-pottiaceous genus as outgroup. With Ptychomitrium as outgroup (Cladogram 10), Timmiella was found to be the most primitive pottiaceous genus, and this genus was also at the base of the tree with Polytrichum (Cladogram 7) and Timmia (Cladogram 9) as outgroups. Timmiella (Pottiaceae) was then used (Cladograms 13–16) as functional outgroup to avoid the possibility that Ptychomitrium was less distant from the Pottiaceae than Aloina (cf. Cladograms 1, 4 and 8) or Gertrudiella, and to avoid problems introduced by reduction in Ptychomitrium.

Cladograms 1–4 and 7–10 support the use of Timmiella as functional outgroup in Cladograms 13–16. They likewise show that at least some genera of Pottiaceae retain characteristics of the more primitive mosses Polytrichum and Timmia. Also, putative sister group genera (Bryobartramia, Ceratodon, Diphyscium, Ditrichum, Encalypta, Grimmia, Syrrhopodon) in the Pottiales, Dicranales, and Grimmiales are mostly too modified in morphology to serve as outgroups in that too many characters are lacking or fixed in states deemed here to be the result of evolutionary reduction. These cladograms demonstrate that many genera are found in similar or even identical lineages in cladograms generated with different outgroups, implying evolutionary development of quite distinctive synapomorphies and supporting more firmly the present hypothesis of relationship.

Cladograms 11 and 12 are from data sets restricted to terminal taxa of two major branches of Cladograms 9 (outgroup Timmia) and 10 (outgroup Ptychomitrium) that are identical between cladograms. These consensus trees are based on a known number of equally parsimonious trees and generally support details of branching patterns in the other cladograms, for which only a portion of all trees could be kept in computer memory and thus contribute to the consensus tree.

The cladogram that presents the hypothesis used here for suprageneric classification is the single tree shown in Cladogram 16. Character state changes for this tree are detailed in Cladogram 15. This single tree was chosen from more than 1250 equally most-parsimonious trees as summarized in the strict consensus tree in Cladogram 14. The major subclades of Cladogram 16 are identified as the subfamilies and tribes recognized in this treatment.

INTRODUCTION:

Outgroup Selection

The argument for past emphasis on sporophyte characters in distinguishing genera of the Pottiaceae is (1) such characters work well in distinguishing taxa in other groups, and (2) taxa sharing similar sporophytes in the Pottiaceae also share similar gametophytes. The present study of the genera of the Pottiaceae has shown, however, that although the general morphology of the well developed capsule and its peristome is similar and typically pottiaceous (with notable exceptions, e.g. Leptodontiella, Streptotrichum, Trachycarpidium and relatives), there is considerable variation in expression of individual sporophyte features, i.e. in degree rather than in kind. Also, detailed morphological analyses demonstrate that this variation is often considerable among taxa with gametophytes that share many morphological and anatomical features not previously evaluated across the family (e.g. the genera Hennediella, Tortula and Weissia as emended here). Arguments for special weighting of sporophytic characters (cf. Crosby 1974, Dixon 1932, Miller 1979) in the Musci generally are not supported at the generic level in the Pottiaceae.

The rationale used here for selecting an outgroup with shared presumed primitive characters is that haplolepideous sister groups to the Pottiaceae (e.g. Calymperaceae, Ditrichaceae, Encalyptaceae, Grimmiaceae) are evolutionarily much reduced and simplified in both gametophytic and sporophytic morphology. This is true when such groups are compared to the range of morphotypes seen in the Pottiaceae. It is necessary to examine the other haplolepideous and even diplolepideous and nematodontous taxa for genera retaining characters of the ancestral morphology. The “generalized” ancestor is not necessarily a gametophytically relatively characterless taxon (e.g. Grimmia), but may be character-rich.

Ptychomitrium (Ptychomitriaceae, haplolepideous) was found to be the closest sister group to the Pottiaceae that was not so morphologically reduced as to have important plesiomorphic character states (also present in more distant ancestors Timmia and Polytrichum as will be demonstrated below) much modified or completely eliminated. This genus is considered the primary outgroup, and was used to establish a functional outgroup among the pottiaceous genera; cladograms were generated with additional alternate non-pottiaceous outgroups in an attempt to evaluate character state polarizations more globally.

Although its peristome is nematodontous, Polytrichum is considered here an acceptable outgroup because it shares many characters with the Pottiaceae that are not associated with apparent reduction. Polytrichum is evidently a moss with very primitive characters, sharing with ferns, for instance, a leaf trace connecting the leaf hydroid strand and the stem central strand, and it has been demonstrated cladistically to be of a lineage more primitive than the Bryales (discussion of Mishler & Churchill 1984). Polytrichum is also distant from the Pottiaceae because of (1) the several unique characters of its peristome, which presuppose a considerable lineage involved in their development, and (2) the number of the species in its genus (and family), representing multiple evolutionary events. The similarity of the gametophytes of several genera of the Pottiaceae with Polytrichum indicates that the Pottiaceae, unlike other families of the Pottiales, retains members with primitive morphologies. Characters of Polytrichum that are, on the other hand, typical of reduced members of the Pottiaceae include epapillose upper laminal cells, lamellae present on the ventral surface of the costa, and annulus vesiculose.

The heterolepideous genus Timmia (Timmiaceae, diplolepideous) is also a major source of information on plesiomorphic features of the Pottiaceae. This genus of the monotypic family Timmiaceae is extraordinarily similar in gametophytic characters to Timmiella, Gertrudiella and other large and presumably non-reduced genera of the Pottiaceae. Like Polytrichum, Timmia has a wealth of characters, all of which are found in the Pottiaceae except the following: inclined capsule position, stomates occurring in several rows in the lower half of the capsule, exothecial cells with sinuous walls, operculum short-mammillate, outer peristome is present, and inner peristome, although similar to that of the Pottiaceae in being filamentous, with 64 segments, and the Primary Peristomial Layer to Inner Peristomial Layer cell ratio (cf. Edwards 1979, Shaw et al. 1989) is apparently that of the Diplolepideae rather than the Haplolepideae. Timmia shares the diplolepideous type of peristome with other families of non-pottiaceous taxa, and would not be expected to appear high in the Pottiaceae tree. Vitt (1984) treats Timmia as quite a derived taxon in the Bryales.

Timmia, Timmiella and Gertrudiella are also quite like Polytrichum in range and development of gametophytic characters; Polytrichum is possibly farther removed from the Pottiaceae than Timmia because of its nematodontous peristome, but it must be recognized that the sharing of numerous gametophytic characters generally not found in similar combination elsewhere in the mosses is evidence of a close phylogenetic relationship between these haplolepideous, diplolepideous and nematodontous genera. The weighting of the three peristome types depends (or should depend) on the number of characters involved in each and is not pursued to any extent in this work. Consideration should be given to the possibility that haplolepideous families may have evolved more than once, and that past weighting of certain characters of the peristome (e.g. presence or absence of the outer peristome) should be eliminated. Shaw et al. (1989) found that although peristome developmental data unite haplolepideous mosses, there is as yet no information as to whether or not this is a synapomorphic condition or not.

Timmia also has characters associated with reduced members of the Pottiaceae, including stem sclerodermis absent, leaves tubulose above, costa elliptical in transverse section, sometimes monoicous, and spores rather large. It is here considered that the gametophytes of Polytrichum, Ptychomitrium and Timmia represent the character states of a non-reduced ancestor of the Pottiaceae better than do other Pottiales. The possibility that the Pottiaceae may be been derived from a reduced immediate ancestor (e.g. other Pottiales) is less probable because the character state combinations of the many Pottiaceous taxa of large stature, which share so many characters with Polytrichum and Timmia, must then have been derived independently. The nematodontous peristome of Polytrichum and its relatives in the Polytrichaceae may well have been derived from arthrodontous and secondarily eperistomate ancestors with gametophytes similar to that of Timmia.

Again, outgroups for the phylogenetic analysis were not selected from only apparently close sister groups, e.g. other Pottiales such as Calymperaceae or Encalyptaceae, or from the Grimmiaceae (see Churchill 1981), because of the potential masking effects of extensive reduction in these groups, at least compared to taxa of the Pottiaceae. This is clearly demonstrated in cladograms including genera of non-pottiaceous families (Cladograms 1–4) in which some non-pottiaceous genera appear, not at the bottom of the tree, but at or near the ends of branches, among what are here considered advanced pottiaceous taxa. Cladograms 5 and 6, with Encalypta and Grimmia respectively as outgroups, place Timmia and Polytrichum together at the end of an apparently highly evolved subclade comprising the here-accepted (Cladograms 14–16) basal stem of the Pottiaceae. This would require re-evolution of a large number of character states that in combination phenocopy the gametophytes of two unrelated non-pottiaceous taxa.

The essential characters of many non-pottiaceous sister genera are those of reduction, and are therefore likely to result in convergence. If this is the case, the primitive members of such groups are unknown, and may well not be as similar to the stem genera of Pottiaceae as are Ptychomitrium, Polytrichum and Timmia, which appear near the base of the trees.

Thus, if Polytrichum and Timmia are both distant from the Pottiaceae and from each other because (1) of the number of unique characters in their peristomes, (2) the number of species (major evolutionary events) in their genera, (3) the fact that other genera share the non-pottiaceous characteristics, and (4) they share more traits with the Pottiaceae than other mosses, then they should appear at the base of the Pottiaceae cladogram. This assumes that the Pottiaceae is monophyletic, of which the twisted peristome found in various subclades is evidence. Of the eight non-pottiaceous genera (other than Polytrichum and Timmia) that were used as outgroups, only Ptychomitrium forced Polytrichum and Timmia low in the tree. The other seven non-pottiaceous genera are considered, therefore, too modified (probably by morphological reduction and simplification) to act as outgroups in calculating a hypothetical phylogenetic tree.

Other Comments

Although two species of Ditrichum (Ditrichaceae), D. tortipes (Mitt.) Par. and D. ambiguum Best, have the twisted peristomes otherwise unique to the Pottiaceae, the gametophytes of that genus are much reduced, and a case might be made for deriving the genus from ancestors of Barbula sect. Hydrogonium. The cladograms, however, do not support for this (but see Cladogram 4).

Encalypta (Encalyptaceae), is variously placed in the Haplolepideae or Diplolepideae (see Edwards 1979 and Vitt 1984). Encalypta is apparently a closer sister group than Timmia because of the quasi-haplolepideous peristome, but there is considerable modification of characters, including monoicy, loss (compared to Timmia) of stem central strand (in most species), gain of yellow KOH reaction of upper laminal cells, loss of costal ventral stereid band, and elaboration of additional layers of costal guide cells. Bryobartramia (Encalyptaceae, see treatment of excluded taxa) is a taxon probably reduced from Encalypta-like ancestors (note that this study will suggest that monotypic genera of reduced morphology are probably relicts of once larger and more complex genera now all but extinct), demonstrating convergence in traits associated with reduction in the Pottiaceae: protonema persistent, small size of gametophyte; very short seta; spherical and cleistocarpous capsule; large spores; and papillose calyptra (that of Encalypta is occasionally papillose).

Autapomorphic characters (e.g. the hyaline exothecial cells of Uleobryum) were not included in the data set in that they are of no value in determining relationship.