Cabochonicus, a new genus for species of gemmate megaspores previously referred to Verrutriletes

Several species of Mesozoic megaspores that are sculptured with globular, often gemmate elements are sufficiently distinct from most verrucate forms to warrant their inclusion in a genus other than Verrutriletes. Cabochonicus is erected to accommodate them with C. (al. Verrutriletes) carbunculus as the type species. The form of the gemmae, which have a glossy appearance, has been the subject of past debate with one interpretation being that they are the manifestation of fungal attack rather than sculpture. Evidence from ultrastructural studies suggests, however, that they are exinal excrescences. The spores are assumed to be the products of heterosporous lycopods and the structure of the wall layers may indicate selaginellalean affinities. The plants are more likely to have been terrestrial than aquatic or semi-aquatic.


INTRODUCTION
In an attempt to determine whether the spherular surface elements of the Mesozoic megaspore Verrutrifetes curbunculus and similar forms are morphological features or the product of some external phenomenon, several specimens attributable to the genus were thin sectioned and examined under a transmission electron microscope (TEM). A t the same time, it was hoped that the structure of the rest of the protective sporopollenin wall layers could be ascertained, thus enabling comparisons to be made with the composition of exines of modern megaspores which, in turn, could possibly shed some light on the botanical affinity of the extinct plants that produced them.
Two main conclusions were reached concerning the first of these aims: 1) the gemmae are surface excrescences that can be accepted for descriptive purposes as sculptural elements and 2) neither Verrutriletes van der Hammen ex PotoniC nor Trifeites Erdtman ex PotoniC accommodate the spores satisfactorily. The arguments in favour of setting up the new genus that is necessary are presented here using nomenclature that has been appropriate to date. This leads to a short systematic section in which Cubochonicus is formally erected and species attributable to it are listed. The paper concludes with a brief consideration of the botanical affinity of the genus and of the palaeoecological implications of its distribution.

BACKGROUND TO PROBLEM
The generic name Verrutriletes was erected by van der Hammen (1954) but it was a nomen nudum until 1956 when a type species was designated by PotoniC. He selected one of the megaspores that Dijkstra (1949) had described from the Upper Cretaceous of the Netherlands, namely Triletes compositipunctutus. This is a relatively small spore (usually significantly less than 500pm in diameter) with conspicuous triradiate ridges and a surface sculpture that consists of numerous granula, verrucae and rugulae. These were originally described as being hemispherical elements 5-20pm high, 5-30pm broad, sometimes smaller (5pm) on the contact areas and often united into star-shaped groups or rows (70pm long) on the distal surface (Dijkstra, 1949;PotoniC, 1956).
PotoniC's diagnosis of the genus allows for the surface elements to range from hemispherical verrucae to low coni of variable size, and for the proximal face to lack sculpture. Binda & Srivastava (1968) emended it to include specific reference to the trilete rays which they stated may or may not reach the equator, PotoniC not having allowed for the former condition.
Seven species were originally included in the genus. There are now more than 30, approximately half of which have been described from Cretaceous rocks; all are Mesozoic. The taxonomic status of a few, and in particular of Verrutriletes curburiculus (Dijkstra, 1949) PotoniC, 1956 has, however, become increasingly confused during the past few years because of uncertainty about the true nature of the sculptural elements which are not strictly verrucae. Dijkstra (1949). described them as hemispherical, red, translucent objects. Poto-niC (1956) noted that they are irregularly scattered like glossy, resin droplets; this led him to place the species only provisionally in Verrutriletes. Knobloch (1984a, pp. 168-169;1984b, p. 156) has referred to them as "glass beads dyed red-brown" and 'pearl-like orna-ments", and for one specimen he compared them to small "bits of tar". Marcinkiewicz (1979) decided that the spherules are not sculpture but fungal elements which are attached to the surface of the essentially smooth-walled Trileites murrayi (Harris, 1961) Marcinkiewicz, 1971. They were placed in a new genus and species, Reymanella globosa. Other megaspores whose exines she interpreted to be similarly modified by fungal organisms are Triletes imitatus Dijkstra, 1961(transferred to Verrutriletes by Waksmundzka, 1982, Verrutriletes pseudosquamosus Marcinkiewicz, 1962 andVerrutriletes sp. in Marcinkiewicz (1962). Verrutriletes guttatus Marcinkiewicz, 1971 can also be included in this list.
Besides the obvious bead-like appearance of the elements, there are a few additional points in favour of Marcinkiewicz's argument. The spherules are often basally constricted and appear to be held onto the spore by means of short appendages. Banerji et al. (1984) noted that they dissolve in acid without leaving a trace on the surface of the spores they examined (recovered from probable Lower Cretaceous rocks in Kachchh, India); regardless of the type of acid used (not stated in their paper), this suggests a composition different from that of the protective wall. The common sedimentary association of V. carbunculus with unornamented but otherwise similar Trileites murrayi could be taken to indicate that they are the same taxon, the latter not having been affected by R. globosa.
Several arguments can, however, also be made against such an interpretation. The spherules look far more like "resinous" droplets or excrescences than fungal bodies. 'lhey are commonly concentrated in a broad band centred on the equator (e.g. Heuber, 1982, pl. 2, figs. 4, 5 ;Kampmann, 1983, pl. 7, figs. la-d). In some specimens, the entire distal surface is also covered (e.g. Knobloch, 1984b, pl. 1, fig. 2). This suggests a morphological character and not a fungal infestation, which would be expected to produce a random pattern. Apparently haphazard distributions do occur on some specimens, and they may be present in patches on the proximal face between and/or on the triradiate ridges (e.g. Marcinkiewicz, 1980, pl. 3, figs. 2, 3). On the other hand, Hueber (1982, pl. 2, fig. l), Kampmann (1983, pl. 7, fig. lc) and others have illustrated specimens whose original tetrad configuration is clearly delineated by smooth contact areas bordered by closely-spaced spherules.

MATERIAL AND METHODS
Verrutriletes (Cabochonicus in systematic section and on Pls. 1-4) carbunculus is usually recovered only as isolated specimens in preparations for large plant microfossils. Those reported on herein come from Lower Cretaceous (Wealden) samples, and from Dijkstra's type assemblage which is of Senonian age. The ornamentation of both sets of specimens varies from consisting of a few widely scattered, relatively large globular elements to smaller, more densely arranged and less well rounded bodies whose bases frequently merge together.
By comparison with the majority of Cretaceous megaspores, V. carbunculus is large (sometimes more than lmm in diameter) and usually brittle, whole specimens being uncommon. The Lower Cretaceous       collection includes some complete spores but with the exception of one comparatively small form, the type material is entirely fragmentary, consisting of fewer than ten additional bits. In view of both the scarcity of undamaged specimens and the large size of the spores, only small fragments were thin sectioned. The procedure adopted for this was as follows: each piece was 1) immersed in a saturated solution.of Uranyl acetate for 30 minutes, 2) washed three or four times in distilled water, 3) soaked in a 1:l mixture of EMix epoxy resin and Analar Acetone, and 4) embedded in resin formulated to set to a medium hardness. The resin blocks were trimmed and the thin sections cut using glass knives and Reichert OmU3 and OmU4 ultramicrotomes respectively. Selected sections were placed on Formvar-coated slot grids, examined under a Philips 301 TEM and photographed. The scanning election-micrographs were obtained using microscopes at the University of East Anglia and the British Museum (Natural History). The specimens were coated with gold prior to examination.

SURFACE MORPHOLOGY
A typical representative of Verrutriletes carbunculus in polar view shows well rounded to globular bodies scattered over the surface of the spore in equatorial and subequatorial regions, the triradiate ridge and areas between the rays being devoid of them (Pl. 1, fig. 3; referred to Cabochonicus). These elements mostly appear to be held in place by "straps" or "root-like'' structures extending from their bases onto the surface of the exine (Pl. 1, figs. 1, 2; P1. 2, figs. 1,2). Although at low magnifications the latter looks smooth, at high I power it is seen to be uneven and irregularly perforated (PI. 1, figs. 2, 4). Without the spherules the species is clearly similar to some smooth walled species, particularly Trileites murrayi (Harris, 1961) Marcinkiewicz, 1971 (PI. 1, figs. 5 , 6). Marcinkiewicz (1979, pl. 3, fig. 3) figured an almost identical specimen which she identified as T. murrayi with spherules of Reymanella globosa Marcinkiewicz on its surface. Other closely comparable forms are illustrated in Pelzer & Reigel (1982, pl. 1, figs. 5 , 6), Huckriede (1982, pl. 8, figs. la-c, 2), and Collinson et al. (1985, fig . 3b). The last of these authors chose not to follow Dijkstra (1951) and positively referred their Lower Cretaceous specimens to Verrutriletes carbunculus. Most of the other illustrations of the species in the literature are also very similar although some differences in density, size, shape and distribution of the spherules is evident. Several show a tendency for the ornament to be concentrated, or to comprise larger elements, in equatorial and subequatorial regions (e.g. in Hueber, 1982;Kampmann, 1983;Knobloch, 1984a, b;Banerji et al., 1984;Batten et al., 1984;Batten & Li, in press).
Verrutriletes imitatus (Dijkstra, 1961) Waksmundzka, 1982 is a similar spore. It differs in supposedly having a less well rounded outline in polar view and in being more evenly covered with spherules. In practice, its distinction from V . carbunculus is not always clear. V . pseudosquamosus Marcinkiewicz, 1962 is stated to be characterised by the occurrence on its outer surface of "somewhat transparent" discs (Marcinkiewicz, 1962, p. 491;1979). The globular bodies on V . guttatus Marcinkiewicz, 1971 are very variable in size, some being well over five times as large as others. Trifetes gamerroi Baldoni & Taylor, 1985 is similar to Verrutriletes sp. in Marcinkiewicz (1962) in bearing only isolated spherules; both of these are like sparsely ornamented V . carbunculus.        The other two specimens illustrated here by scanning electron micrographs (Pl. 1, figs. 7-9) are not readily referable to any of these species. They are more densely ornamented, and the elements are smaller, less globular and frequently merge together at their bases. One of them (Pl. 1,figs. 8,9) is also sculptured on the contact areas and along the triradiate ridge.

ULTRASTRUCTURE
A number of terms have been applied to the layers that make up megaspore walls. Several are synonyms, their use depending on author preference but one, perine, has been the subject of controversy. Kempf (1970Kempf ( ,1973 argued vigorously that this word should be employed for the thick spongy outer wall, with exine being reserved for the much thinner inner stratum. Although such an interpretation i s accepted by some (e.g. Bergad, 1978), many authors regard both layers as exine (e.g. Pettitt, 1966;Tryon & Lugardon, 1978). They are labelled exoexine and intexine herein.
Transmission electron micrographs of thin sections of the two walls in Verrutriletes carbunculus indicate that the bulk of the exoexine consists of ramifying sporopollenin threads of varying thickness which are orientated approximately normal to the outer surface of the spore, whereas the intexine is a thin, lamellated layer. The exoexinal threads are of constant form except adjacent to the inner surface of the outer wall where they become smaller and/or more densely distributed (e.g. P1. 2, figs. 7, 8) and in the vicinity of the sculptural elements where they may merge to form more robust units (e.g. PI. 2, fig. 5). By contrast, the exoexine of a specimen of Trileites murrayi recovered from the Middle Jurassic of Bornholm by Kempf (1971, and referred to therein as Istisporites murrayi (Harris) Kempf, 1971) is made up of three zones of differing character. The inner wall is, however, comparably thin and of similar composition to that in V. carbunculus.
The exoexine of specimens 1 and 3 (Pl. 2, figs. 3-8; P1. 3, figs. 1-3) from Dijkstra's collection comprises a particularly open structure with the ratio between threads and intervening air-space being approximately 1:1. On the other hand, the rodlets in specimens 2 (Pl. 3, figs. 4-7) and 4 (PI. 4, figs. 1-6) are more densely distributed; indeed in the former they are mainly tightly packed. It is possible that the open meshwork reflects some expansion of the wall following maceration in Schulze solution and immersion in Potassium hydroxide. Dijkstra (1949, p. 21) noted that at first he prepared his samples in this way, but later found soaking in water was all that was required for the release of spores from the inorganic matrix. Specimen 2 may not, therefore, have been treated with these chemicals and specimen 4 certainly was not.
The intexine of the thin sectioned spores proved to be less than 1/20th the thickness of the.exoexine. It is composed of sporopollenin membranes of up to 0.1pm thick arranged parallel to the surface forming a dense lamellar structure (Pl. 3, fig. 2). The junction with the exoexine is clearly delineated within a narrow zone (maximum width c. 0.4pm) where the outermost laminae are seen to be discontinuous and partly interspersed with exoexinal threads. Thus, where the two layers are separated, a little of the outer wall commonly remains attached to the inner (PI. 3, fig. 3).

DISCUSSION
The observations above on the surface morphology of Verrutriletes carbunculus spores agree with those that have been made by Marcinkiewicz (1979Marcinkiewicz ( , 1980, Knobloch (1984a) and Waksmundzka (1985). The last of these authors also demonstrated that the exoexine consists of irregularly massed sporopollenin threads that pass upwards into dense "verrucae". Her thin

Explanation of Plate 3
All ultrathin sections of fragments of Cabochonicus carbunculus from Dijkstra's (1949) type collection.       section illustrations (Waksmundzka, 1985 , pl. 14, figs. 1-5) tend to emphasise the degree of separation of the surface elements as do several of those herein (e.g. P1. 2, fig. 3; PI. 4, fig. 5) but there can be much broader areas of contact with the outer surface (e.g. P1. 2, figs. 4, 8; P1. 4, figs. 3, 4, 6). Knobloch (1984a) pointed out that a secretion of a viscous fluid would necessarily produce basal attachment processes. In some instances, the spherules themselves bear excrescences (e.g. Hueber, 1982, pl. 1, fig. 3). Hueber (ibid.) noted that they appear to have been exuded from within the spore as a resinous substance which retained its glossy surface and viscous appearance after solidifying. The internal foam-like composition of the spherules on his Lower Potomac (Virginia) specimens contrasts, however, with the largely solid form of both the Polish (Waksmundzka, 1985) and west European material. Similar structuring may be apparent within degraded elements if the spongy exoexine is raised up under them in the manner shown on P1. 4, figs. 3, 4. It could also be a corrosion effect.
Although looking as if they have been exuded from within the spore, the thin sections do not indicate any exoexinal modification that supports such an interpretation. Despite their shiny appearance in reflected light, there is no evidence of a compositional difference but rather merely a gradation from a spongy texture to solid masses. Their mode of formation thus remains problematical. Whatever their origin, however, the fact that they are not true verrucae renders unsatisfactory the inclusion in Verrutriletes of spores on which they occur. Since thin sectioning has proved that the elements should be treated as a form of sculpture and not the product of fungal attack, reference to Trileites is also inappropriate. The only logical course of action to follow is, therefore, to erect a new genus to accommodate them.

Diagnosis:
A triradiate megaspore of simple form that bears glossy, resinous-looking gemmae and other sculptural elements surmounting an otherwise essentially smooth exoexine. These surface features tend to be concentrated in equatorial and subequatorial regions and/or on the distal surface but they may also occur on the proximal face between and on the rays of the triradiate ridge. They can be widely scattered or densely packed with bases merging together, and may be as small as granules or comprise large (verrucasized) spherules and protruberances. Description of holotype: Dijkstra did not designate a holotype for Triletes carbunculus but the single specimen he figured (Dijkstra, 1949, pl. 2, fig. 12) can be accepted as such. It is preserved, albeit now in a fragmented state, in a slide labelled "Holotype" along with a much smaller but complete specimen and a few other pieces. The size and shape of the fragments indicate that it was originally subspherical. Comparison with his figure suggests that it was c. 940pm in diameter including the sculpture (c. 900pm without).  Marcinkiewicz, 1971, p. 56, p1. 5, figs. 4-8;pl. 6, figs.   form (all ~2 5 0 0 ) . scattered gemmae. These elements are not mentioned in either the diagnosis or the description of the species because they were considered to be "foreign bodies". It is tentatively suggested here that reference to Cabochonicus rather than Triletes might be appropriate for at least part of the assemblage on which this taxon is based. Stratigraphic range of genus. Published records suggest that this is Late Rhaetian to Santonian.
Remarks. Hueber (1982, p. 5 ) noted that his spores "seem encrusted over most of their surfaces with minute, lustrous, ruby cabochons". Reference of these elements to polished but uncut precious stones seems even more appropriate than other previously employed qualifiers such as glass beads and resinous spherules, hence its selection as the basis of the new generic name. Truly verrucate spores should continue to be referred to Verrutriletes as originally described by PotoniC (1956) and emended by Binda & Srivastava (1968). Trileites is maintained for essentially smooth-walled forms. It is accepted that some specimens referable to the latter genus may be in fact be degraded examples of Cabochonicus which have lost their excrescences but this can be difficult to prove. Corrosion often removes characters that are critical for spore identification. It is not a problem unique to these genera.
Cabochonicus cannot be satisfactorily classified using Potonie's (1956 et. seq.) scheme since it is neither a smooth nor a verrucate spore and hence is not referable to either of the Infraturmas Laevigati or Apiculati. No attempt is made to place it elsewhere or to create a new category because this is not seen to serve any useful purpose. The value of purely morphological classification is becoming increasingly unclear as research continues on dispersed spores and their relationships with the plants that produced them.

AFFINITY
It is assumed that Cabochonicus is a lycopod megaspore. In common with others of the group, the sporopollenin wall layers comprise a thick, spongy exoexine and a much thinner intexine. Of the two modern heterosporous orders, the Selaginellales and the Isoetales, the external form and internal structure of Cabochonicus bear more resemblance to the products of a few species of Selaginella (cf. Minaki, 1985) than to fsoetes but no direct comparison is possible. The exines of spores of extant fsoetes and of fossil forms that have been attributed to the Isoetales (e.g. Hall, 1963;Pettitt, 1966;Collinson et al., 1985;Kovach & Dilcher, 1985) are more open in texture and no laminated intexine is detectable. Origins within now extinct families such as the Pleuromeiaceae are possible but positive evidence is wanting.
The selaginellids have a long history that extends from the Carboniferous and living members, of which there are about 700, show considerable ecological adaptation in habitats. They are, however, usually terrestrial. The tendency for Cabochonicus to occur only sporadically and in low numbers in sediments suggests that the plants which produced them also inhabited dry land. This is in sharp contrast to species of such genera as Minerisporites PotoniC 1956 and Puxillitriletes Hall & Nicolson, 1973 which may be very abundant and suggest local derivation from aquatic or semi-aquatic members of the Isoetaceae (Batten, 1975(Batten, , 1982Collinson et ul., 1985;Kovach & Dilcher, 1985).

CONCLUSION
The natural affinity of the heterosporous plant that produced Cabochonicus carbunculus and other species of the genus remains uncertain but a subgroup of the Selaginellales is a possibility. The sculpture is, however, clearly neither verrucate nor the result of fungal attack but rather consists of exinal excrescences that are often gemmate in form.