Systematics

Phylum Bryozoa Ehrenberg, 1831 Order Cheilostomata Busk, 1852 Suborder Flustrina Smitt, 1868 Infraorder Neocheilostomina d'Hondt, 1985 Family Doryporellidae fam. nov.

Diagnosis: Anascan-grade neocheilostomes with well-developed cryptocystal frontal wall ornamented with a reticulate pattern of polygonal ridges. Gymnocyst greatly reduced. Opesia bell-shaped to oval. Oral spines present, branched or unbranched, one to three pairs adjacent to opesia and sometimes additional spines. Avicularia present, adventitious. Ovicell hyperstomial, ectooecium uncalcified, endooecium reticulated. Pore chambers present. Ancestrula tati-form or modified tatiform.

Type genus: Doryporella Norman, 1903; family also includes Doryporellina Grischenko, Mawatari and Taylor, 2000.

Remarks: The phylogenetic analysis described below unites Doryporella and Doryporellina as a clade. The previous inclusion of these two genera within the Calloporidae is unsatisfactory because this family is diverse and paraphyletic. We therefore propose the new family Doryporellidae which can be distinguished from Calloporidae principally by the well-developed polygonal reticulation of the cryptocyst of the autozooids.

Genus Doryporella Norman, 1903 Doryporella smirnovi sp. nov. (Figs 1A–F and 2A–D)

Fig. 1A–F

Doryporella smirnovi sp. nov., ZIRAS 1/50130. A, general view of the holotype colony; B, colony growing margin with developing zooids; C, mature zooids with spherical ovicells (some broken) and associated avicularia; D, mature zooids, showing details of ovicells and avicularia; E, immature zooid showing bell-shaped opesia, oral spines and lateral avicularium; F, frontal shield surface ornamented with polygonal reticulation. Scale bars, 1mm (A), 0.5 mm (B, C, D), 0.2 mm (E), 0.05 mm (F).

Fig. 2A–D

Doryporella smirnovi sp. nov., ZIRAS 1/50130. A, paired lateral avicularia; B, proximal margin of opesia with groove and bicus-pate ooecial fold of developing ovicell; C, ancestrular region; D, ancestrula with spines on peripheral gymnocyst and tubercles around opesia (distal end of ancestrula is towards bottom left). Scale bars, 0.1 mm (A, B), 1mm (C), 0.2 mm (D).

Doryporella alcicornis: Grischenko, 1997: 160; Grischenko et al., 2000: 251 (part).

Material examined: Holotype: ZIRAS 1/50130, mature colony, about 6×5 mm in size, encrusting a pebble. Pacific Institution of Bio-Organic Chemistry Collection on the R.V. Akademik Oparin (14th Expedition), 24 August 1991, Stn 47, Bering Sea, sector to north-west of Bering Island (Commander Islands archipelago), 55°25.2′N, 165°33.8′E, depth 152 m, sand, Trawl Sygsbi; collector A.V. Smirnov.

Diagnosis: Doryporella with bell-shaped opesia; 4–5 oral spines; usually two avicularia along one lateral edge of each zooid; ancestrula modified tatiform with dimpled cryptocyst and tubercles around edge of opesia.

Etymology: The new species is named in honour of Dr Alexei V. Smirnov (Zoological Institution Russian Academy of Science, St Petersburg) who collected the material.

Description: Colony encrusting, multiserial, sheet-like (Fig. 1A), more or less circular, attaining 6 mm maximum width, light-yellow in colour when dry. Autozooids elongate, hexagonal to irregularly oval (0.52–0.93 mm long × 0.42– 0.51 mm wide), separated by distinct narrow sutures (Fig. 1C). Vertical gymnocystal walls with a pair of distal and two pairs of distolateral pore chamber windows (Fig. 1B). Frontal wall cryptocystal, convex, non-porous, ornamented by a pattern of polygonal reticulate ridges, most obvious in young zooids, non-granular. Opesia bell-shaped with straight proximal edge, slightly wider than long (0.12–0.18 mm long × 0.15–0.22 mm wide); a shallow crescentic groove extends along the outside of the distal margin, and the frontal wall proximal of the opesia lacking depressions. Operculum shape as opesia but smaller, with hemispherical thickening. Oral spines hollow, up to 0.19 mm long, four to five in number (Fig. 1C–E): two to three spines at the distolateral corners of the opesia; and a pair located a little more proximally. Oral spines orientated upwards and distally. Avicularia adventitious, located along lateral margins of zooids (0.07–0.13 mm long), usually a pair on one side (Fig. 2A), sometimes symmetrically placed near opposite proximolateral corners of the opesia (Fig. 1B). Avicularia occasionally single or absent. Mandible subtriangular, pointed, without calcified pivotal bar, orientated laterally or proximo-laterally. Ovicell (Fig. 1A–D) prominent, spherical (0.20–0.25 mm long × 0.20–0.25 mm wide), with surface having same reticulate pattern as frontal wall, as long as broad, ectooecium uncalcified. Additional avicularium associated with ovicells (0.05–0.08 mm long), located along distal gymnocystal margin of ovicell, rostrum directed distally.

Ancestrula (Fig. 2D) modified tatiform, small (0.25 mm long × 0.21 mm wide), with finely dimpled cryptocyst. Six short spines associated with narrow peripheral gymnocyst. Opesia subcircular (0.09 mm long × 0.11 mm wide), with smooth margin having seven short bulge-like tubercles along lateral and proximolateral border. Two distolateral periancestrular zooids budded from the ancestrula. Periancestrular zooids (Fig. 2C) small (0.33 mm long × 0.21 mm wide), wide, oval, with inflated frontal walls, with or without single avicularium. Three pairs of oral spines present in some zooids from primary zone of astogenetic change.

Remarks: Doryporella smirnovi sp. nov. differs from other species of Doryporella in the absence of modified sub-oral or asymmetrically arranged spines. The type species of the genus, D. spathulifera, possesses a large flattened sub-oral spine associated with a median suboral avicularium which is not developed in D. smirnovi. Two other species, D. armata and D. alcicornis, both have large, oval opesiae whereas this structure is smaller and bell-shaped in D. smirnovi. Additionally, among species of Doryporella only D. smirnovi has paired lateral avicularia and a modified tatiform ancestrula with a wrinkled cryptocyst.

Distribution: The new species is presently known only from the area of the Commander Islands (Bering Sea) and can be categorized as a Pacific High-Boreal species.

Phylogenetic analysis

The discovery of the new species of Doryporella described here has forced a re-evaluation of the phylogeny of this genus and its relationship with Doryporellina. In an earlier paper (Grischenko et al., 2000) revising the then known species of Doryporella, a cladistic anlaysis was performed based on skeletal morphological characters. Twenty-one skeletal characters were coded for the four putative species of Doryporella along with five out-group species of calloporid cheilostomes chosen because of their morphological and biogeographical similarities to the species of Doryporella. After removal of one of the out-group species (Megapora ringens (Busk, 1856)) whose position proved to be unstable, a single most parsimonious tree was recovered using the phylogenetics program PAUP 3.1. The tree grouped together three of the four putative species of Doryporella as a clade, but excluded from this clade Doryporella reticulata Ryland, 1963, showing it instead as being more closely-related to Amphiblestrum flemingii (Busk, 1854). These results led to the proposal of the new genus Doryporellina for D. reticulata in order to maintain the monophyly of Doryporella. The reticulated cryptocystal frontal shield present in both Doryporella and Doryporellina was interpreted by Grischenko et al. (2000) to have evolved convergently.

A new phylogenetic analysis incorporating the recently-discovered species Doryporella smirnovi was performed to establish the correct generic affinities of this species (Doryporella or Doryporellina), and to ascertain the effect on the phylogenetic tree of adding this new taxon. The same out-group species were used as those employed by Grischenko et al. (2000), with Megapora ringens again being excluded from the final analysis. The original 21 characters were all coded and three extra characters were added: presence of an avicularium associated with the ovicell, granulation of the cryptocyst, and reticulation of the ovicell roof (Table 1). The analysis was undertaken using PAUP 4.0 after first confirming that the original results of Grischenko et al. (2000) were replicated using the same character matrix in this updated version of the program. Minimum length trees were found using the branch and bound option.

Table 1.

Character matrix for species of Doryporella and Dory-porellina and out-group taxa employed in the cladistic analysis. Characters 1–21 are the same as those used by Grischenko et al. (2000); characters 22–24 are novel to this new analysis.

Analysis of the in-group taxa alone yielded an unrooted tree of 24 steps with CI 0.875 (0.700 when uninformative characters are excluded) and RI 0.571. Two clades occur in this tree: smirnovi + reticulata, and armata + spathulifera. Bootstrap analysis of 1000 replicates found the first of these clades in 65% of the trees and the second in 64%. The armata + spathulifera clade was also found in the earlier analysis of Grischenko et al. (2000) and the new data corroborates a sister group relationship between these two species of Doryporella. Addition of the out-group taxa yielded 4 equally parsimonious trees of length 41 steps and having a CI of 0.683 (0.629 when uninformative characters are excluded), and RI of 0.649. In an exhaustive search of 135135 random trees generated from the data set, mean tree length was found to be 57.07 (SD 4.27), g1= 0.476552 and g2 = −0.185902. Among these random trees 37 were only one step longer than the most parsimonius trees, 158 two steps longer, 334 three steps longer and 550 four steps longer. A majority rule concensus tree computed from the four equally parsimonius trees corresponded to the topology of tree number 2. This favoured tree is shown in Fig. 3. Once again there is a clade comprising armata + spathulifera but the smirnovi + reticulata clade found in the in-group analysis is absent. All four equally parsimonious trees place smirnovi as the sister taxon to the three established species of Doryporella, while three of the four trees have Doryporellina reticulata as the sister taxon to this Doryporella clade, the other tree grouping D. reticulata with Amphiblestrum flemingii as in the original analysis of Grischenko et al. (2000).

Fig. 3.

Favoured tree, one of four equally parsimonious trees of 41 steps, rooted on Tegella unicornis, showing the inferred relationships between species of Doryporella, Doryporellina and out-group taxa. Note that the new species Doryporella smirnovi forms the sister group of the three established species of Doryporella, and that Doryporellina reticulata is the sister group to the four species of Doryporella. The new family Doryporellidae comprises the clade Doryporella + Doryporellina. Tree statistics: CI = 0.683 (0.629 when uninformative characters are excluded); RI = 0.649. Apomorphies, with character numbers in brackets, are as follows: Branch 1: (13) suboral avicularium, (14) lateral avicularium; Branch 2: (1) extensive, shelf-like cryptocyst, (4) one distal pore chamber, (5) two disto-lateral pore chambers, (6) bell-shaped opesia, (20) ovicell longer than wide; (22) no avicularium associated with ovicell; Branch 3: (15) proximal avicularium; Branch 4: (19) opesia wider than long, (21) autozooid less than 1.5 × longer than wide; Branch 5: (5) more than 2 distolateral pore windows; Branch 6: (6) semielliptical opesia; Branch 7: (7) beaded opesia rim; (19) opesia longer than wide; Branch 8: (2) reticulation of cryptocyst, (24) reticulated ovicell; Branch 9: (3) beaded cryptocyst, (7) beaded opesia rim, (11) separated outer pair of oral spines, (15) proximal avicularium, (16) calcified pivotal bar of avicularium, (18) closure plates, (20) ovicell wider than long; Branch 10: (8) groove distal of opesia, (9) branched oral spines, (14) lateral avicularium, (22) avicularium associated with ovicell, (23) ungranulated cryptocyst; Branch 11: (4) two distal pore windows; Branch 12: (6) ovoidal opesia, (17) tatiform ancestrula, (19) opesia longer than wide; Branch 13: (12) suboral spine/s, (13) suboral avicularium; Branch 14: (4) two distal pore windows, (6) bell-shaped opesia; Branch 15: (5) one distolateral pore window, (21) autozooids about 2 × longer than wide; Branch 16: (10) asymmetrical distribution of oral spines, (22) no avicularium associated with ovicell.

Comparison of this new cladistic analysis with that of Grischenko et al. (2000) underscores the volatility of cheilostome phylogenies based on a relatively small number of skeletal morphological characters. It is clear that analyses incorporating soft part anatomy, and more particularly using molecular phylogenies, are strongly needed in cheilostomes.

Taxonomic implications

Initial uncertainty regarding the generic attribution of the new species described in this paper is dispelled by the phylogenetic analysis which shows that it is a sister taxon to the three species forming the Doryporella clade. Therefore, assignment of the new species to Doryporella is supported by the analysis.

In contrast to the findings of Grischenko et al. (2000), the new analysis supports a close relationship between Doryporella and Doryporellina. Three of the four most equally parsimonious trees place Doryporellina in a sister group relationship with Doryporella. The original argument for erecting the genus Doryporellina is undermined by these new results because inclusion of the species reticulata in Doryporella no longer renders the genus polyphyletic. Instead, the reticulate frontal shield that was doubtless the reason why Ryland (1963) attributed his species reticulata to Doryporella is, according to the new analysis, an apomorphy that evolved only once in this group of Arctic-Boreal calloporid species. (Note that evidence still remains for the independent acquisition of similar reticulate frontal shields in other cheilostome bryozoans, e.g., Biflustra reticulata Tilbrook et al., 2001). While it would now be possible to place Doryporellina in synonymy with Doryporella, this is not mandatory because Doryporellina lies outside the clade formed by the other species of Doryporella and therefore its existence does not render Doryporella paraphyletic. The morphological distinctiveness of Doryporellina argues for the retention of the genus. Among the characters of Doryporellina permitting its distinction from Doryporella are the beads on the opesial rim and on the ridges forming the reticulations of the cryptocyst, the separated outer pair of oral spines, and the occasional occurrence of proximal adventitious avicularia with calcified pivotal bars and of closure plates occluding the opesia.

The Calloporidae as currently understood is a very large family of cheilostomes – Gordon (unpublished) lists 74 genera in this family. Among the Neocheilostomatida (i.e., those cheilostomes having brooded coronate larvae), the Calloporidae are both primitive and paraphyletic (Gordon, 2000). The oldest neocheilostome is the calloporid Wilbertopora Cheetham, 1954 from the mid-Cretaceous of the USA, and it is likely that most or all other neocheilostome families nest within the Calloporidae. Clearly, the Calloporidae requires splitting if it is to become a taxon of phylogenetic value. One way of achieving this is by eliminating from the family any well-characterised monophyletic group comprising two or more genera. Doryporella is a highly distinctive genus and its classification as an anascan belonging to the Calloporidae has been questioned in the past (Levinsen, 1916) and also by D.P. Gordon in his unpublished listing of cheilostome genera for the Treatise on Invertebrate Paleontology. With the recognition of a clade comprising Doryporella + Doryporellina there are grounds for a new family-level taxon, the Doryporellidae as diagnosed above and distinguished principally by the extensive, reticulated cryptocyst.