Peristome Development in Mosses in Relation to Systematics and Evolution

Peristome Development in Mosses in Relation to Systematics and Evolution. II. Tetraphis pellucida (Tetraphidaceae) Author(s): Jonathan Shaw and Lewis E. Anderson Source: American Journal of Botany, Vol. 75, No. 7 (Jul., 1988), pp. 1019-1032 Published by: Botanical Society of America Stable URL: http://www.jstor.org/stable/2443770 Accessed: 11/11/2010 06:37

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=botsam.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Botanical Society of America is collaborating with JSTOR to digitize, preserve and extend access to American Journal of Botany.

http://www.jstor.org Amer. J. Bot. 75(7): 1019-1032. 1988.

PERISTOME DEVELOPMENT IN MOSSES IN RELATION TO SYSTEMATICS AND EVOLUTION. II. TETRAPHIS PELLUCIDA (TETRAPHIDACEAE)1

JONATHAN SHAW2 AND LEWIS E. ANDERSON Departmentof Botany, Duke University, Durham, North Carolina27706

ABSTRACT The Tetraphidaeis a small subclass of mosses with a nematodontous peristome that has frequentlybeen interpretedas primitiveamong the truemosses. The developmentalcell sequence leading to the formationof the four peristometeeth of Tetraphispellucida is describedfor the firsttime. Comparisonsare made with sequencesknown for other nematodontousand arthrodontous mosses. Peristomedevelopment in T. pellucidais more like that describedpreviously for arthrodontousperistomes than to published developmentalsequences for nematodontous peristomesof species in the Polytrichaceae.On the other hand, our observationsconfirm a basic uniformityof the earliestdevelopmental stages in all mosses studiedthus far,regardless of their systematicposition.

THE TETRAPHIDAEis the smallest subclass of phidaceaerepresents a primitive trait retained mosses, consistingof a singlefamily, the Tetra- from an ancestor in common among the Te- phidaceae Schimp., and two genera, Tetraphis traphidae, Sphagnidae,and Andreaeidae. Hedw. and TetrodontiumSchwaegr. Tetraphis It is the structureof the peristome that has is comprisedoftwo species, T.pellucida Hedw. been most frequently cited as evidence for a and T. geniculata Girg. ex Milde, and Tetro- primitive phylogenetic position for the Tetra- dontium is a monotypic genus erected for T. phidae. The peristome in both Tetraphisand brownianum (Dicks.) Schwaegr. All three Tetrodontiumconsists of four massive teeth species of the Tetraphidaeare north-temperate that fill the opercularregion of the maturecap- in distribution. sule. The teeth are nematodontous in that they Although there are no serious taxonomic are made up of whole, thick-walledcells, rather problemsat the specificlevel within the Tetra- than just remnants of cell walls, as in arthro- phidae, the phylogeneticand taxonomic status dontous peristomes. of the subclass has been disputed through the Moss peristomes can be classified as arthro- years. Many bryologists have expressed the dontous or nematodontous on the basis of view that the Tetraphidaeretain a significant comparative morphology without necessarily number of features that are primitive in the making hypotheses about the phylogenetic re- true mosses. The protonema of species in the lationshipbetween the two types ofperistomes. Tetraphidaceaehas thalloid "flaps" that are When Philibert (1884) distinguished arthro- often compared to the thalloid protonema of dontous and nernatodontous peristomes, he Sphagnum L. and Andreaea Hedw. (e.g., compared them in terms of cellular structure Campbell, 1905; Goebel, 1905; Schofield, and the patterns of lines on their inner and 1985). Although homology between the thal- outer surfaces.The morphological distinction loid protonema in these groups has not been is a phenetic and not a phylogenetic one, al- explicitly hypothesized, the suggestion, or at though Philibert and many other bryologists least the possibility, of homology is implicit in since have attached phylogenetic significance the frequentcomparisons that have been made. to the differencesin morphology. It is, however, still an open question as to Fundamentalsimilarities in the basic ground whether the thalloid protonema of the Tetra- plan among superficiallydivergent arthrodontous peristomes have led many bryologists to the interpretationthat mosses having arthro- 'Received for publication 3 June 1987; revision ac- dontous peristomesare monophyletic. That is, cepted 1 October 1987. basic similarities suggest to many that the ar- 2 Present address:Department of Biology, Ithaca Col- throdontous peristome has evolved only once lege, Ithaca, NY 14850. in the mosses. These fundamentalfeatures in- This researchwas supportedby NSF Grant No. BSR- 8506992. We thankB. D. Mishlerfor assistancewith field clude the formation of teeth from adjacent work and for commentingon an earliermanuscript draft. periclinalwalls ratherthan whole cells, the ar- We also thank M. Tumer for his able technicalassistance. ticulated appearance resulting from the pat-

1019 1020 AMERICAN JOURNAL OF BOTANY [Vol. 75 tems of lines on the inner and outer surfaces, *stometeeth in this species to published se- and the uniform number of cells in the outer quencesfor species in other subclassesof moss- and primaryperistomial layers throughoutar- es. In particular,we have sought to answerthe throdontousmosses. Anatomical studies have following questions during the course of our shown that almost all arthrodontous mosses study: 1) Is therea patternof development (i.e., have 32 cells in the outer peristomial layer a sequence of cell divisions) that is unique to (OPL)and 16 in the primaryperistomial layer and shared by the nematodontous peristomes (PPL) (Edwards, 1984; Shaw and Robinson, of species in the Polytrichidae and Tetraphi- 1984). Moreover, developmental studies, al- dae? 2) At what stage of development do Po- though rather few, indicate a remarkableuni- lytrichum-and Tetraphis-typeperistomes di- formity of development among arthrodontous verge? 3) At what stage of development do peristomes (Evans and Hooker, 1913; Tetraphis-typeand arthrodontousperistomes Blomquist and Robertson, 1941; Saito and diverge? In order to make comparisons be- Shimoze, 1954; Saito, 1956). tween peristomedevelopment in Tetraphisand Edwards(1979) devised a convention, which that in other mosses, we will rely on published he termed the peristomial formula, to make reports by others, as well as on our own pub- comparisonsof the number of cells in the peri- lished (Shaw, Anderson, and Mishler, 1987) stomial layers of mosses easier. The formula and unpublished observations. A detailed re- lists the number of cells in each of the three view of peristome structureand development, layers, from the outside inward, in only one intended to serve as a backgroundand ratio- eighthof the capsule'scircumference. The peri- nale for this and our future contributions,was stomial formula of almost all arthrodontous provided in Shaw et al. (1987). mosses ranges from 4:2:2 to 4:2:14 (Edwards, 1984; Shaw and Robinson, 1984). That is, the MATERIALSAND METHODS- Tetraphis pel- OPL (outer peristomial layer) uniformly con- lucida is widespread across the boreal and sists of 32 cells and the PPL (primaryperisto- ,north-temperate regions of the Northern mial layer)of 16 cells, but the inner peristomial Hemisphere. In North America, it extends layer (IPL) ranges from 16 to 112 cells in cir- northwardinto the tundra of Alaska and the cumference.Most of this variation in the num- Yukon Territory,and southwardin the East to ber of cells in the IPL occurs between species; South Carolinaand Alabama and in the West within a given species the number of cells in to Arizona and California,where it occurs in all three layers is relatively constant. the mountains-(Crum and Anderson, 1981). A monophyleticorigin for groupshaving ne- Populations sampled for this study originated matodontous peristomes is less clear. If ne- in the Ap-palachianMountains of North Car- matodontousperistomes are primitive with re- olina, where T. pellucida is common -on rotten spect to arthrodontous peristomes, as many tree -stumps, logs, and on soil or rocks near bryologists assert (e.g., Cavers, 1911; Dixon, streams. Collection data for the populations 1924; Edwards, 1984; Schofield, 1985), then we sampled are as follows: Mitchell Co.: Roan the nematodontous structureitself cannot be Mt., Anderson and Mishler 24,533; Anderson evoked as evidence for monophyly. It would and O-Toole 24,71 . Voucher specimens are simply be a feature that is retained from a deposited in DUKE. Plants were collected dur- common ancestor. Furthermore,the two sub- ing autumn and midwinter when very small classes of mosses which are characterizedby sporophytes (<4 mm long) were visible am.ong nematodontous peristome structure (Poly- the perichaetialleaves. Sporophytesat this stage trichidaeand Tetraphidae) have peristomesthat were fixed immediately in formalin-acetic-al- are extremely differentin cellularground plan, cohol, and the remaining plants with young in spite of sharingthe feature of being formed sporophytes were allowed to mature in the lab- from whole, persistent cells. Gametophytic oratory. Plants were kept slightly moist in cov- structurein the two subclasses also suggests, .ered dishes and were sampled periodically for because of extreme differences,that the Poly- sporophytes of different developmental stages. trichidae and Tetraphidaemay not be closely Sporophytic tissue was embedded in paraffin related (e.g., Crum and Anderson, 1981). and sectioned on a-rotarymicrotome according In this paper we describe the sequence of to the standard techniques (Johansen, 1940). cell divisions in the peristomial region of de- For additional details regarding our methods veloping sporophytes of Tetraphispellucida. of fixation, embedding, and sectioning, see This is the first developmental study of pei- Shaw et al. (1987). stome structure in any species of the"Tetra- phidae, and we will compare the sequence of RESULTS-The sporophyte of T. pellucida divisions preceding deposition of the peri- grows in length as a result of divisions of a two- July 1988] SHAW AND ANDERSON-PERISTOME IN TETRAPHIS 1021 sided apical cell, as in other mosses. However, eight cells surroundingfour inner cells (Fig. 7, as also seems to be the case in other mosses 8). The eight outer cells are the amphithecium which-we presentlyhave under study, the api- and the four inner cells, the so-called funda- cal cell of T. pellucidaceases dividing at a.very mental square, are the endothecium. early stage of sporophytic development. We The sequence of cell divisions described up found no evidence of a functioning apical cell to this point representsserial views of the de- in sporophytes longer.than approximately 3- veloping sporophytefrom the apex downward 5 mm, a develop.mentalstage. at whichthe young through progressivelyolder portions. The re- spear was barely exerted from the perichaetial maining descriptionsthat follow, however, de- leaves. After the apical cell ceases dividing, the scribe the proliferation and differentiationof apex of the sporophyte becomes broadly the embryonic cell layers at the level of dehis- rounded and divisions occur in the youngest cence of the operculum (i.e., the "annular" merophytesjust below the tip (Fig. 1). These level- Tetraphishas no specializedannulus per observations.suggest that most elongation of se). Development as seen from transversesec- the sporophyteof T. pellucidaoccurs as a result tions will be emphasizedsince majorperistome of enlargementand division of cells -belowthe types in mosses are distinguished mainly on -apex,but we did not determine the manner in the basis of the numbers and arrangementof which the setae develop. Since the maturespo- cells in the peristome-forminglayers. rophyteof T. pellucidais often 15-20 mm long, Soon after the endothecium and amphithe- only about one-third of this length is the direct cium have differentiated,divisions begin in the result of divisions of the apical cell. amphithecium(Fig. 8). Anticlinaldivisions oc- While the apical cell is still functioning, the cur in each of the eight amphithecialcells, and, distal portion of the sporophyte is relatively more or less synchronously, periclinal divi- sharply 'pointed (Fig.-2).. The first.transverse sions occur, initiating the formation of a two- section closest.to. the apex of such a.young layered amphithecium (Fig. 8). In general, the sporophytesometimes shows a singlecell which periclinaldivisions occur first,resulting in two is the apical cell .itself (Fig. 3). At a: slightly amphitheciallayers, each ofwhich is composed lower-level (Fig.-4),; a transverse view shows of eight cells. Then anticlinal divisions occur two-cells, one of which is probably the apical in the outer of these two layers, resultingin an cell and the otheris the firstmerophyte. Slightly outer layer of 16 cells and an inner layer of lower still (Fig. 5),. three cells are visible;-the eight cells. However, the sequence of divisions apical cell is in the center, flanked on either is not always so regular,nor are all of the an- side by the two most recently formed and still ticlinal or periclinaldivisions in each layer ab- undivided merophytes. Sections of any given solutely synchronous. Figure 8, in fact, shows sporophyte do not always show each of these severalamphithecial cells dividing anticlinally arrangementsofocells. Sometimes, for example, before they have divided periclinally. In this the most distal-sectionobtained from the young figure,two other cells have divided periclinally sporophyte-isbelow the level at which the api- before any anticlinal divisions have occurred; cal cell alone is visible in transverse view. In the latter representsthe typical order of divi- other cases, the orientation of the walls of the sions. first two merophytes is such that both deriv- As a result of these anticlinal and periclinal atives plus the apical cell cannot be viewed in divisions in the first and second amphithecial a single transverse section like that shown in layers, whether the divisions are absolutely Fig. 5. In many cases, the arrangementof cells synchronousor not, eventually two layers are typically seen in the first transverse-secti-on in formed (Fig. 9). At this stage there are almost a series from apex to base is as in Fig. 4. invariably 16 and 8 cells in the outer and inner Each of the most recently formed mero- layers,respectively. The inner of these two lay- phytes first undergoes an anticlinal division ers is the inner peristomiallayer, or IPL. Next, resulting in four cells in tranverse view (Fig. periclinal divisions occur in the outer layer, 6). Shortly thereafter,the first stage in the dif- producinga three-layeredamphithecium (Fig. ferentiation of the two embryonic layers, en- 10).The middle layer,with 16 cells at this stage, dothecium and amphithecium, commences is the primary peristomial layer, or PPL. The with the formation of a series of curved anti- peristomial formula at this stage is 0:2:1; the clinal walls (Fig. 7). One wall is formed in each divisions that form the OPL have not yet oc- cell shown in the previous quadrant (Fig. 6), curred. producinga transient stage in which the spo- Almost as soon as the amphitheciumis three- rophyteconsists of eightcells in transverseview layered, cells in the outermost layer undergo (Fig. 7). Almost immediately, curved periclinal anticlinaldivisions (Fig. 10-many of the cells walls are formed, resultingin an outer layer of in the right half of the section have divided; 1022 AMERICANJOURNAL OF BOTANY [Vol. 75

1 i 2

* * 1

Fig. 1-6. Transverseand longitudinalsections throughyoung sporophytesof Tetraphispellucida. (Fig. 1, 2 x 400; 3-6 x 648). 1. Longitudinalsection of a sporophytein which the apical cell has ceased dividing. 2. Longitudinalsection of an embryonicsporophyte in which the apical cell is still dividing. The most recently formed merophyteis on the left; the second merophyteis on the right.3. Transversesection throughthe apical cell. 4.Transversesection fromjust below the apex of a sporophyteas in Fig. 2, showing two cells. 5. Transversesection from just below the apex of a sporophyteas in Fig. 2. The apical cell is in the center, flankedby the first and second merophyts. 6. Transverse section of a young sporophyteslightly lower than Fig. 4, 5. Each merophytehas divided anticlinally,producing four cells. The calyptrasurrounds th'e sporophyte. those in the left half have not). When these completed,anticlinal divisions in the IPL com- anticlinal divisions have been completed, the mence (Fig. 1 1). Each IPL cell divides roughly outer of the three amphithecial layers consists in half, increasingthe number of cells in this of 32 cells (Fig. 11). layer to 16 (Fig. 1 1-13). Each new anticlinal At about the same time that anticlinal di- wall in the IPL is offset slightly with regardto visions in the outer layer have increased the correspondinganticlinal walls in the PPL, this number of cells in this layer to 32, and some- resultingin one of the IPL cells being slightly times slightly before the divisions have been largerthan the other of each pair. The degree July 19881 SHAW AND ANDERSON-PERISTOME IN TETRAPHIS 1023

to which the new walls are not aligned with we will refer to this most recently derived in- walls in the PPL is somewhat variable, but the nermost amphithecial layer as the IPL-2. The positions seen in Fig. 11 are typical. alternative, calling it the IPL, would effectly Divisions in the four endothecial cells also resultin a displacementofthe threeperistomial commence at this stage (Fig. 11). The early layersinward, and would make it virtuallyim- pattern of divisions in the endothecium is al- possible to make interpretations about ho- most exactly like the sequence of divisions mologies betweenthe peristomiallayers in Tet- leading to the original differentiation of en- raphis with those in other mosses. dotheciumand amphithecium.Each ofthe four At about the same time that cells in the IPL endothecial cells first undergoes an anticlinal, are enlarging,anticlinal divisions occur in the or tangential,division (Fig. 12), followed by a PPL (Fig. 14). Such divisions do not occur in periclinaldivision. The resultis an innersquare every cell of the PPL, and, in fact, they occur of four cells, surroundedby an outer layer in in roughly every other PPL cell. As a result, which additional anticlinal divisions occur al- where there were previously two PPL cells ad- most immediately so that this layer consists of jacent to four OPL cells in each eighth of the 12 cells (Fig. 13). This outer endothecial layer capsule's circumference,the number of PPL gives rise to the sporogenoustissue in the cap- cells increases to three. The resulting peristo- sule urn. Few additional divisions occur in the mial formula at this stage is thus 4:3:2 (Fig. endotheciumin the opercularregion ofthe cap- 14). sule and even at a relatively mature stage the Because of the differentiationof the IPL-2, operculum contains relatively few, conspicu- alongwith additionalpericlinal divisions in the ously large, cells (Fig. 22). outermost amphithecial layer, the amphithe- At approximately the same time that anti- cium becomes six to seven layers in thickness clinal divisions occur in the IPL, periclinaldi- (Fig. 15, 16). Also, anticlinal walls in the am- visions begin in the outer layer, producing a phithecium become noticeably less aligned in four-layeredamphithecium. When the fourth the concentric layers than they were in earlier amphitheciallayer has been completelyformed, developmental stages, and as a result, it be- the three peristomial layers, OPL, PPL, and comes more difficult to specify a peristomial IPL, are differentiated,and the peristomialfor- formula. Additional anticlinal divisions occur mula is 4:2:2 (Fig. 12, 13). in the PPL so that the formula becomes 4:4:2, Up to this point in development, the ar- at least in some areas in the circumferenceof rangement of cells in the amphithecial layers the sporophyte(Fig. 16). Divisions in the IPL is highly organized,and anticlinal walls in the also increase the number of cells in this layer three layers are for the most part aligned with and result in a formula of roughly 4:3-4:3-4 one another. However, as development pro- (Fig. 16). Over 20 sporophytes of T. pellucida gresses,involving increasesin both the number were sectioned in this study; variation in the and size of cells, transverse sections reveal a numberof primaryand inner peristomiallayer decreasingdegree of orderlinessin the arrange- cells in one eighth of the capsule's circumfer- ment of cells (e.g., Fig. 13-16). ence occurred both within and between spo- Periclinal divisions occur in the outer am- rophytes. However, the formula was consis- phithecial layer so that the 4-layered amphi- tently 4:3-4:3-4. That is, the number of PPL thecium becomes 5-layered (Fig. 13, 14). The and IPL cells range from 24 to 32, but do not periclinaldivisions in the outer layer are even seem to vary beyond these limits. less synchronous than the earlier divisions in While cells in the amphithecial layers are the amphitheciallayers, and typically any giv- undergoinganticlinal and periclinaldivisions, en section shows some, but not all, cells of the as described above, the cells in each layer are outer layer divided (e.g., Fig. 14). Also at this roughly isodiametric to short-rectangularas stage, cells in the IPL, and to a lesser extent, seen in longitudinalview (Fig. 17). At this stage the PPL, begin to enlarge (Fig. 14). The IPL six or seven or occasionallyfewer layers of cells cells then undergo periclinal divisions, again make up the amphithecium at the level of de- in a somewhat disorganizedand asynchronous hiscence of the operculum, which, in Fig. 17, fashion, to producean additionalamphithecial is towardthe base ofthe photograph.The spore layer adjacentto the endothecium (Fig. 15). It layer, which is the outermost layer of the en- is noteworthythat all previous increasesin the dothecium, is clearlyrecognizable in the latter number of amphithecial layers had been the figure.Its cells stand out because they are more result of periclinaldivisions in cells of the out- heavily stained and have thickerwalls than the ermost layer. In orderto maintain a consistent adjacentcells. The spore layer can thus be used application of the terms, IPL, PPL, and OPL, as a markerto distinguish amphithecium and 1024 AMERICAN JOURNAL OF BOTANY [Vol. 75

endothecium. It should be noted that the open gitudinal section through the upper portion of space in the lower right of the section in Fig. the urn and the lower part of the operculum 17 is a tear and is not an air space. The capsules in which cell elongation and enlargement are of Tetraphis lack an air space. complete, and secondary wall deposition has The next two layers outside the spore layer begun. The line-of opercular dehiscence is near represent IPL- 1 and IPL-2, which originated the upper end of the photograph. The spore from periclinal divisions of the original IPL, layer is visible at the lower portion of the figure, as described above (Fig. 15, 16). The next or identifiable by the dark-staining, compact, third layer of cells toward the outside in- Fig. square to short-rectangular cells which end 17 is the PPL, beyond which are the OPL-1 considerably short of the ring of dehiscence of and -2. The outermost layer of the amphithe- the operculum. Although there has been some cium becomes the capsule wall. tearing in sectioning in Fig. 18, the individual Shortly after the stage shown in Fig. 17, all peristomial layers can be identified by relating of the cells within the operculum begin a period them to the spore sac, which marks the outer of dramatic elongation, enlargement, and dif- layer of the endothecium. ferentiation, as a result of which the capsule Higher magnification of longitudinal sec- assumes its mature shape. The cells of the out- tions through the peristomial region after elon- ermost peristomial layers, including the OPL- 1 gation and considerable wall deposition reveals and -2 and the PPL, undergo a striking elon- .clearly the cellular details of the developing gation with little or no increase in width. This peristome, shown in Fig. 1.9, 20. The line- of results in long and narrow cells that dovetail opercular dehiscence in Fig. 19 is near the bot- past each other, forming prosenchymatous-like tom, while in Fig. 20 it is near the top. The tissue (Fig. 18-22). The cells of both IPL- 1 and line can be recognized by a series of contrasting -2 and all of the cells of the endothecium be- narrower cells in the opercular wall. The cells come elongated, considerably widened, and ul- of the three outer- peristomial layers. (OPL- 1 timately are emptied of their contents (Fig. 18- and -2 and the PPL) in these longitudinal views 22). Wall deposition begins in all of the cells are long, narrow, and'.pointed, while the cells of the peristomial layers immediately after the of the IPL- 1 and -2 are much wider and are period of elongation (Fig. 18) and continues more or less rectangular. Apparently only the through maturation. Figure 18 shows a lon- outer wall of the IPL-.2 receives secondary

Fig. 7-12. Transversesections of sporophytesat progressivelyolder developmentalstages. (x 648). Calyptraeare visible in Fig. 7, 8. 7. Differentiationof the endothecium and amphitheciumby a series of curved anticlinal and periclinalwalls. 8. Four endothecialcells surroundedb-y the amphitheciumin which periclinaland anticlinaldivisions are occurring.9. Two-layeredamphithecium; the innerlayer consists of 8 cells, the outerof 16 cells. The fourendothecial cells form the so-called fundamentalsquare and have not yet divided. 10. Three-layeredamphithecium surrounding the four endothecial cells. Anticlinal divisions are occurringin the outer amphitheciallayer. 11. Three-layeredam- phitheciumwith anticlinaldivisions in the outerlayer completed; there are now 32 cells in this layer.Anticlinal divisions are occurringin the inner layer. Note that the anticlinalwalls of the IPL are not quite alignedwith anticlinalwalls in the PPL. The endothecialcells are beginningto divide. 12. Periclinaldivisions are occurringin the outer amphithecial layer, producingthe OPL. Anticlinal divisions are occurringin the IPL and a second cell of the endothecium has divided.

Fig. 13-18. Transverseand longitudinalsections throughsporophytes of T. pellucida. (Fig. 13, 14, 17 x648; 15, 16, 18 x 400). 13.Four-layeredamphithecium surrounding the endothecium which has now undergonerepeated divisions. Note that the arrangementof cells in the amphitheciallayers is not as organizedas in earlier sections. The peristomial formula is 4:2:2. 14. Periclinal divisions in the outer layer have produced the fifth amphitheciallayer. Anticlinal divisions in the PPL have resulted in a 4:3:2 peristomial formula. 15. Continuingbut irregularanticlinal divisions in the outermost amphitheciallayer. IPL cells are swelling and undergoingpericlinal divisions, producing the IPL-2. 16. The IPL-2 is completelydifferentiated; the amphitheciumnow consists of seven layers. The peristomial formulais still 4:3:2. 17. Longitudinalsection througha young sporophyte.Cells in all the endothecialand amphithecial layersare isodiametricto short-rectangular.18. A somewhatolder sporophytein which the cells in the opercularregion have elongated.

Fig. 19-24. Transverseand longitudinal sections through developing sporophytes.19, 20. Longitudinalsections throughprogressively older sporophytes.Cells in the outer amphitheciallayers are narrowlylinear and the walls are beginningto thicken. (x 648). 21, 22. Longitudinalsections throughimmature (Fig. 21 x 100) and nearly mature(Fig. 22 x 160) sporophytes.Note theoabsence of air spaces in the urn in Fig. 22. 23. Transversesection of a nearlymature sporophyteat approximatelythe level of the capsulemouth. ( x 648). 24. Transversesection of a nearlymature sporophyte at approximately1/3 the way up the opercularregion. Note wall thickeningsin endothecialcells. (x 648). July 19881 SHAW AND ANDERSON-PERISTOME IN TETRAPHIS 1025

97 10;_

lo/

11 'slv12 1026 AMERICANJOURNAL OF BOTANY [Vol. 75

\ 13 14

;s~~~~~~~~~~~~~~~~~. V

* . . S S.~~I

,1If', July 1988] SHAW AND ANDERSON PERISTOME IN TETRAPHIS 1027

19 20

21'( ... .-4

?z','^ 1028 AMERICANJOURNAL OF BOTANY [Vol. 75

`265

Fig. 25-27. Transversesections throughnearly mature capsules in the opercularregion and an SEM of the mature peristome.25. Transversesection throughthe opercularregion in the upperthird. (x 648). 26. Transversesection from near the apex of the opercularregion. Note wall thickeningsin endothecialcells in 25, 26. (x 648). 27. SEM of the matureperistome. (x 64). thickening (Fig. 19, 20), but there is clear evi- cavity. Fartherbelow, not shown in the figure, dence from SEM studies of mature peristomes the columella persists. Final wall deposition in (Fig. 27) that these broad rectangularcells of the peristomial cells is not complete until well the IPL-2 become the innermost cells of the past meiosis and the individual sporesare freed peristome teeth. These cells show clearly on from the tetrad. the inner surfacesof the teeth in Fig. 27. Cross sections through the operculum of Lower magnifications of longitudinal sec- nearly mature capsules reveal the cellular tions of the entire operculumafter the peristo- structureand organizationof the peristome of mial cells are differentiatedshow clearly the Tetraphis(Fig. 23-26). The cross section shown relationships of the various layers whose de- in Fig. 23 is near the dehiscence line of the velopmental sequences we have described operculum.The large,empty, thin-walledcells above. Figure 21 is not quite median, but the in the centerof the section are mostly endothe- approximatelyfive peristomiallayers are clear- cial cells. The outermost layer of largercells is ly distinguishable.The outermostofthe broad, presumed to be the IPL-2, judging from its rectangular, clear cells comprise the IPL-2, position in longitudinalsections. Mostly, only while the three columns of cells in the center the outer walls of this layer are thickened. In are endothecial.(The centralcolumn is narrow contrast to the precision and regularityfound because the section is not median.) in the peristomiallayers of arthrodontousperi- A still lower magnificationis shown in Fig. stomes, the peristome of Tetraphisshows con- 22, in which the upperportion of the spore sac siderablevariability. The alignment of cells in is apparent. The entire upper portion of the the various layers in Fig. 23 is not precise, columella has lysed and the sporocytes, now indicating that there have been anomalous in meiosis, lie free in the relatively large spore periclinaland anticlinal divisions. It is not al- July 19881 SHAW AND ANDERSON -PERISTOME IN TETRAPHIS 1029

ways possible to distinguishbetween endothe- atodon purpureus (Hedw.) Brid., Aulacom- cium and amphithecium. nium heterosticum (Hedw.) B.S.G., Bartramia Figures24 to 26 representcross sections se- crispata Schimp., and Dicranum japonicum lected from points progressively toward the Mitt. (op. cit.). Our unpublished observations apex of the operculum and reflect the conical on approximatelyeight additional arthrodon- natureofthe operculum.There is a progressive tous species support the impression of devel- reduction in the number of endothecial cells opmental uniformity gained from the litera- and it is likely that toward the tip of the oper- ture.This uniformityis all the more impressive culum OPL-2 and IPL-2 layers are lacking. In considering that the developmental observa- Fig. 25, 26 all of the opercularcells are thick- tions have been obtained from a broad range walled and become the terminal portion of the of both haplolepideous and diplolepideous peristome teeth. It is difficultin these positions species, and now also from Tetraphispellucida, in the operculum to distinguish between am- a nematodontousspecies. Data presentlyavail- phithecium and endothecium with certainty, able indicate that development of the peristo- but at least the four centralcells in Fig. 26 and mial layers in these species also shares signif- 27, which are clearly a part of the peristome icant features in common with those of teeth, are endothecial in origin. These sections members of the Polytrichaceae,yet some note- thus demonstrate that in the upper opercular worthy differencesexist as well. These differ- region, endothecial cells become thickened at ences will be discussed below. the time peristomialwall materialis deposited A remarkable feature of peristome devel- in the amphitheciallayers, and close to the apex opment in all mosses that have been studied become a portion of the peristome. is the regularityof the early cell divisions, both At maturity,the tissue formingthe opercular spatially and temporally. The peristomial lay- regionofthe capsulebecomes divided into four ers are clearlydefined concentricrings of cells, massive teeth. Each tooth is composed of the owing to the fact that periclinalwalls of divided thickened, elongated, amphithecial cells (Fig. cells in any given ring are almost perfectly 27). The central thin-walled endothecial cells aligned around the circumferenceof the cap- in the opercularregion break down just prior sule and occur in every cell of the ring. More- to dehiscenceso that the teeth arenot triangular over, both anticlinal and periclinal divisions in transverseview. Rather, both the inner and in any given ring occur more or less synchro- outer faces of each tooth are convex outward nously. In T. pellucida, this synchronization (Fig. 27). The linear shape of the outer cells of decreases as peristome development pro- the peristome teeth can be discerned in the gresses, whereas in arthrodontousperistomes mature peristome; they range from approxi- the divisions are remarkably synchronous -mately 40-100 ,um in length. The teeth are throughoutdevelopment. AlthoughBlomquist approximately700 ,umin total length. and Robertson (1941) provided a somewhat exaggeratedimpression when they described DISCUSSION-Considering that much, if not the divisions in Aulacomnium heterostichum most, of the basis for modern classificationsof as almost absolutely synchronous, it appears the true mosses is based on peristome struc- that even the degree of regularitywe have de- ture, it is remarkablehow little is known about scribedin T. pellucida is highly unusual, if not the developmentofperistome teeth. There have unique, in plant development. been a few thoroughdevelopmental studies of One point of variation among the published arthrodontousperistomes (Evans and Hooker, sequences of peristome development is the 1913; Blomquist and Robertson, 1941; Saito manner in which the amphithecium and en- and Shimoze, 1954; Saito, 1956; Stone, 1961; dothecium are differentiated.Several studies Shaw et al., 1987) and even fewer of nema- indicate that a single periclinal wall is formed todontous peristomes (van der Wijk, 1929; in each cell of a quadrantsuch as that shown Wenderoth, 1931; Chopraand Sharma, 1958, in our Fig. 6. These periclinal divisions dif- 1959; Chopraand Bhandari, 1959). However, ferentiate four inner cells, the endothecium, one prominent feature of the developmental from four outer cells, the amphithecium. An- observationspresently available is the striking ticlinal divisions in the four amphithecialcells uniformity of the earliest sequences of cell di- then result in an inner "fundamentalsquare" visions. Up to at least the point where the of four endothecial cells surroundedby a ring amphitheciumhas become 3-layered(approx- of eight amphithecialcells. A sequencelike this imately Fig. 11), the pattern of development differentiatingthe endothecium and amphi- we have described for Tetraphis is virtually thecium has been reportedfor Funaria hygro- indistinguishable from the patterns reported metrica Hedw. (Kienitz-Gerloff, 1878), Di- for Diphysciumfoliosum (Hedw.) Mohr, Cer- cranum japonicum (Saito, 1956), Bartramia 1030 AMERICAN JOURNAL OF BOTANY [Vol. 75 crispata (Saito and Shimoze, 1954), and Po- In Bartramia crispata (Saito and Shimoze, lytrichumjuniperinum (Wenderoth, 1931). 1954), a diplolepideous species, the first an- Most investigators,however, have observed ticlinal division in each of the eight IPL cells that the endothecium and amphithecium are is nearly symmetrical, resultingin a 4:2:2 cell differentiatedby a series of curved anticlinal arrangement. Each IPL cell then undergoes and periclinalwalls resultingin a pattern sim- anotheranticlinal division, resultingin a 4:2:4 ilar to that shown in our Fig. 7. A pattern of arrangement.The patternin Tetraphisis much differentiationlike this has been reported for like this. Each IPL cell undergoes one nearly Aulacomnium heterostichum(Blomquist and symmetrical anticlinal division, followed by Robertson, 1941), Diphysciumfoliosum(Shaw anticlinal divisions in each of the derivatives. et al., 1987), Polytrichumformosum (Vaizey, In Tetraphis, however, unlike Bartramia cris- 1888), and Lorentziellaimbricata (Rushing and pata, anticlinaldivisions also occurin the PPL, Snider, 1980). resulting in a 4:3-4:3-4 arrangementof cells Chopra and Sharma (1958) reported both in the three peristomial layers. patterns of amphithecial differentiation de- The pattern of development described for scribed above in individual sporophytes. nematodontous peristomes of species in the Moreover, these authors found that variation Polytrichaceaediffers from that of both Tetra- in the mannerof endothecialand amphithecial phis and the arthrodontous species (van der differentiationcan exist between adjacentmer- Wijk, 1929; Wenderoth, 1931; Chopra and ophytes in the young sporophyte.Stone (1961) Sharma, 1958, 1959; Chopra and Bhandari, also found such variation in the manner in 1959). The earliest stages of development are which the amphitheciumand endothecium are similar to those of the arthrodontousspecies differentiatedin Mittenia plumula. Although and Tetraphis,to approximately the point at we have found the patterndescribed above for which the amphithecium is 2-layered. How- Tetraphisto be the most common in the mosses ever, a greaternumber of anticlinal divisions we have under study, we too have observed occurs in the amphitheciallayers of species in both patterns in Ditrichumpallidum (Hedw.) the Polytrichaceae than in arthrodontous Hampe (unpublisheddata). mosses studied to date. In terms of the se- Snider (1975) reporteda very differentpat- quence of cell divisions preceding deposition tern of endothecial and amphithecial differ- of the peristomial wall material, Tetraphisis entiation in ArchidiumBrid. This unique pat- considerablymore similar to those arthrodon- tern of development is sometimes cited as tous species that have been studied than it is evidence suggesting an isolated phylogenetic to species in the Polytrichaceae.With the latter, position for the Archidiaceae (Snider, 1975; Tetraphisshares the nematodontous structure Crum and Anderson, 1981). The two patterns of the mature peristome teeth, i.e., the persis- described for other nematodontous and ar- tence of whole thick-walled cells rather than throdontous species may also have phyloge- only portions ofcell walls. Tetraphiseventually netic significance,although present data indi- has a cellular pattern of approximately4:3-4: cate that the patterns cut across basic and 3-4, and, therefore, has more PPL cells than undoubtedly natural systematic groupingsde- is typical of arthrodontousmosses, but fewer fined by many other morphologicalcharacters. PPL cells than generallyoccur in species in the The development of the peristomial layers Polytrichaceae.Moreover, the increase in the in Tetraphisisvirtually identical to the patterns number of PPL cells occurs in a much later describedfor arthrodontousperistomes at least developmentalstage in Tetraphisthan in species up to the point at which the eight IPL cells of the Polytrichaceae. Evidently, grouping undergo anticlinal divisions (Fig. 12). In Cer- mosses in accordancewith the morphology of atodon purpureus,Dicranum japonicum, and mature teeth (nematodontous vs. arthrodon- Diphyscium foliosum, which are haplolepi- tous) is not congruentwith groupingthem ac- deous species, the first anticlinal divisions in cordingto similar developmental sequences of the IPL are extremely asymetric, producing a cell divisions. Our bias is that developmental larger and a smaller cell (Evans and Hooker, sequences provide a better estimate of ho- 1913; Saito, 1956; Shaw et al., 1987). The next mologies and phylogenetic relationships than anticlinal divisions in the IPL occur in the do comparisonsof maturestructure alone. Sev- largerof these two cells. The resultis three IPL eral bryologistshave previously remarkedthat cells derived from each of the eight original the Polytrichaceaeand Tetraphidaceaedo not cells, producinga peristomialformula of 4:2:3. appear closely related in spite of sharing ne- This cellular pattern, or some slight modifi- matodontous peristome structure (e.g., Lim- cation of it, is considered to be characteristic pricht, 1895; Crum and Anderson, 1981; Ed- of haplolepideousperistomes at maturity(Ed- wards, 1984). wards, 1979, 1984). Many bryologists have expressed the view July 1988] SHAWAND ANDERSON -PERISTOME IN TETRAPHIS 1031 that the Tetraphis-type peristome is primitive CAVERS, F. 1911. The inter-relationships of the Bry- as compared to arthrodontous peristomes ophyta. New Phytol. 9-10: 1-200. (Campbell, 1905; Cavers, 191 1; Dixon, 1924; CHOPRA, R. N., ANDN. N. BHANDARI. 1959. Cyto-morphological studies of the genus Atrichum Palis. Res. Grout, 1936; Crosby, 1980; Schofield, 1985). Bull. Punjab Univ. 10: 221-231. However, except for the unique and peculiar , AND P. D. SHARMA. 1958. Cyto-morphologyof structureof the peristome, consisting of four the genus Pogonatum Palis. Phytomorphology 8: 41- massive teeth, there seems to us to be little 60. objective evidence that the Tetraphis-type , AND . 1959. Cytomorphological studies of peristome is, in fact, primitive. Most authors Oligotrichum Lam. et De Cand. J. Indian Bot. Soc. 38:400-414. who have suggesteda primitive interpretation CROSBY,M. R. 1980. The diversity and relationships of for the Tetraphis-typeperistome, and for the mosses. In R. J. Taylor and A. E. Leviton [eds.], The Tetraphidaceae,give no explicit reason for the mosses of North America, 115-129. Pacific Division, conclusion. Cavers (191 1), for example, stated AAAS, San Francisco. only that "The peculiarand probablyprimitive CRUM, H. A., AND L. E. ANDERSON. 1981. Mosses of peristome of Tetraphismay in some respects North America. Columbia University Press, New be with the of York. compared elaterophore Aneura DIXON, H. N. 1924. The student's handbook of British and Metzgeria... ." Schofield (1985), in his mosses. Ed. 3. V. V. Sumfield, Eastboume. excellent new textbook of bryology, observed EDWARDS, S. R. 1979. Taxonomic implications of cell that "This subclass [the Tetraphidae]is often patterns in haplolepideous moss peristomes. In G. C. considered to be the most generalized of the S. Clarke and J. G. Duckette [eds.], Bryophyte sys- true mosses!" The precise meaning of the un- tematics, 317-346. Academic Press, New York. fortunate . 1984. Homologies and inter-relations of moss term, "generalized,"and its phylo- peristomes. In R. M. Schuster [ed.], New manual of genetic significanceis not, however, at all clear. bryology. Vol. 2, Ch. 12, 658-695. Hattori Botanical Philibert (1889) suggestedthat the Tetraphis- Laboratory, Nichinan. type peristome is a relic, primitive type, and EvANs, A. W., AND H. D. HOOKER,JR. 1913. Devel- that the Polytrichum-typerepresents a modi- opment ofthe peristome in Ceratodonpurpureus. Bull. fication of it. However, none of these authors Torrey Bot. Club. 40: 97-109. provides GOEBEL, K. 1905. Organography of plants especially of any substantive evidence to support the Archegoniatae and Spermaphyta. Part II. Clar- the assumption of primitiveness for the Tetra- endon, Oxford. phis-type peristome, or the Tetraphidae. GROUT,A. J. 1936. Tetraphidaceae. In A. J. Grout [ed.], The idea that the Tetraphidae are actually Moss flora of North America north of Mexico. Vol. more closely related to the arthrodontous 1, 5-7. Newfane, VT. mosses is also not new. Braithwaite(1 887) not- JOHANSEN,D. A. 1940. Plant microtechnique. McGraw- ed that the calyptraof is reminiscent Hill, New York. Tetraphis KIENITz-GERLOFF, F. 1878. Untersuchungenfiber die of those found in species of the Orthotricha- Entwicklungsgeschichte der Laubmoos Kapsel und die ceae, that the leaf areolationis mnioid, but that Embryoentwicklung einiger Polypodiaceen. Bot. Zei- the peristome is unique. Edwards (1984) dis- tung (Berlin) 36: 33-48, 49-64. cussed the possibility that the Tetraphis-type LIMPRICHT,K. L. 1895. Die Laubmoose Deutschlands, peristome is secondarily derived rather than Oesterreichs und der Schweiz. Vol. 2. E. Kummer, primitive. This view is supportedby the sim- Leipzig. PHILIBERT,H. 1884. De l'importance du p6ristome pour ilarity in gametophyticstructure between Tet- les affinities naturelles des mousses. Rev. Bryol. Li- raphis and a number of species in the diplo- chenol. 11: 65-72. lepideous order, Bryales. Although much 1889. Etudes sur le peristome. Rev. Bryol. Li- additionalinformation is needed on peristome chenol. 16: 1-9. development in other arthrodontousand ne- RUSHING, A. E., AND J. A. SNIDER. 1980. Observations on sporophyte development in Lorentziella imbricata matodontous species, our observations (Mitt.) Broth. J. Hattori Bot. Lab. 47: 35-44. strengthen the plausibility of a closer phylo- SAITO, S. 1956. Studies on the development of the peri- genetic relationship between the Tetraphidae stome in Musci. II. On the peristome in Dicranum and arthrodontousgroups than between either japonicum Mitt. Bot. Mag. (Tokyo) 69: 53-58. group and the nematodontous Polytrichidae. , AND S. SHIMOZE. 1954. Studies on the development ofthe peristome in Musci. I. On the peristome in Bartramia crispata Schimp. [In Japanese, English LITERATURE CITED summary.] Bot. Mag. (Tokyo) 68: 55-60. SCHOFIELD,W. B. 1985. Introduction to bryology. Mac- BLOMQuIsT,H. L., AND L. L. ROBERTSON. 1941. The millan, New York. development of the peristome in Aulacomniumhet- SHAW,J., AND H. ROBINSON. 1984. On the development, erostichum.Bull. Torrey Bot. Club. 68: 569-584. evolution, and function of peristomes in mosses. J. BRAITHWAITE,R. 1887. The British moss flora. Vol. 1. Hattori Bot. Lab. 57: 319-335. Acrocarpi I. L. Reeve, London. - L. E. ANDERSON, AND B. D. MISHLER. 1987. Peri- CAMPBELL,D. H. 1905. The structure and development stome development in mosses in relation to system- of the mosses and fems (Archegoniatae). Ed. 1. Mac- atics and evolution. I. Diphysciumfoliosum (Diphys- millan, London. ciaceae). Mem. New York Bot. Gard. 45: 55-70. 1032 AMERICAN JOURNAL OF BOTANY [Vol. 75

SNIDER, J. A. 1975. Sporophytedevelopmentinthegenus WENDEROTH, H. 1931. Beitrage zur kenntnis des Spo- Archidium(Musci). J. Hattori Bot. Lab. 39: 85-104. rophyten von Polytrichum juniperinum Willdenow. STONE, I. G. 1961. The gametophoreand sporophyteof Planta 14: 244-385. Mittenia plumula (Mitt.) Lindb. Austral. J. Bot. 9: VAN DER WIK, R. 1929. Uber den Bau und die Entwick- 124-151. lung der Peristomzahne bei Polytrichum. Recueil Trav. VAIZEY,J. R. 1888. On the anatomy and development Bot. Neerl. 26: 289-394. of the sporogoniumin mosses. J. Linn. Soc. Bot. 24: 262-285.