Image stacking

Stacked photo reconstructions were prepared using a Canon EOS 6D Mark II equipped with either a Canon MP-E 65 mm (lectotype) or 100 mm objective (lectotype and OUMNH specimens). One system (65 mm objective) was fixed on a Cognisys StackShot Macro Rail guided by Helicon Remote software; 30 images were stacked using Zerene Stacker Software. The second system (100 mm objective) involved manual focusing and photo-stacking of four (lectotype) and five (OUMNH specimen) images using open-source software Picolay (www.picolay.de); specimens were blot dried using tissue paper; flash highlights were removed and minor cosmetic corrections were made using the open-source software Paint.net.

Radula reconstruction

An image of the paralectotype radula was reconstructed using an automated Leica DM5500B microscope with motorized Z-focus. Views were captured with a Leica DFC450C (5 Megapixel CCD sensor) digital microscope camera, using Z-Stack acquisition with the multifocus module of Leica Application Suite (LAS: Z-range 565 µm; Z-Step 15 µm; steps 39; with depth of focus determined after visualisation directly through the microscope). LAS software v.4.12.0 was used to fuse different images, and to recombine multiple views using the reconstruction multifocus module. Stitching of isolated stacked views was achieved using Affinity Photo version 1.10.5.

Station Map

RIMS “Investigator” station coordinates were sourced from Anon (1914), and converted from the original degrees, minutes, seconds (Fig. 1, Table 1) into decimal degrees for mapping purposes.

Fig. 1

Type localities, Teretoctopus indicus (closed rhombus) and T. alcocki (cross – lectotype, and large open and closed circles – in Bay of Bengal). For coordinates (latitudes, longitudes), please see Table 1. Inset: RIMS “Investigator” survey path from stations 339–347, with anomalous U-turn from station 342 off Sohar to 343 off Muscat, before a course correction to sample at station 344 within the Strait of Hormuz (closed circle), but more likely in a location depicted by a closed star. Bathymetric contours (light to dark grey): 200, 500, 1,000, 2,000, 10,000 m

Table 1

Recorded (not necessarily correct) coordinates (degrees, minutes, seconds) of Teretoctopus type material: sourced Goodrich (1896), Anon (1914), Robson (1932). IM – Indian Museum, L – lectotype, P – paralectotype, S – syntype; NR – not reported; “?”– denotes uncertainty

Family placement

The Enteroctopodidae Strugnell, Norman, Vecchione, Guzik et Allcock, 2014 was originally differentiated from the Bathypolypodidae Robson, 1929 and Octopodidae d’Orbigny, 1839 on molecular grounds. No recognised morphological criteria consistently differentiate these taxa. Accordingly, without molecular data, our attribution of Teretoctopus to the Enteroctopodidae is provisional, and based on its greater overall similarity to taxa historically attributed to Benthoctopus, Vulcanoctopus, and Muusoctopus than to those of Ameloctopus, Bathypolypus, or Grimpella (not necessarily related genera that share the basic character combination “two rows of suckers along the arms and no ink sac”).

SYSTEMATICS

Order Octopoda Leach, 1817

Suborder Incirrata Grimpe, 1916

Superfamily Octopodoidea d’Orbigny, 1839

Family Enteroctopodidae Strugnell, Norman, Vecchione, Guzik et Allcock, 2014

Genus Teretoctopus Robson, 1929

Type species by original designation, T. indicus Robson, 1929: 608.

Diagnosis. Body smooth; suckers in two rows; ink sac absent; funnel organ reduced, quadripartite; crop with or without anterior diverticulum; radula with lateral cusps on rachidian; posterior salivary gland length < 20% mantle length.

Teretoctopus alcocki Robson, 1932

Figs 1–10, 12–15, Tables 2–4

Teretoctopus alcocki Robson, 1932 – Robson (1932: 252–253, pl. III (fig. 3), text fig. 47); Adam (1939: 105–106, text fig. 25).

?[sic.]Octopus januarii Hoyle, 1885 – Goodrich (1896: 19).

Polypus januarii (Hoyle, 1885) – Massy (1916; partim [specimens 1341/1, 3348/1], 199–200).

Etymology. Though not explicitly stated, this species was most certainly named after Alfred William Alcock, the dual surgeon/naturalist aboard RIMS “Investigator” from 1888 to 1902.

Type material examined. RBINS (I.G. – Inventaire générale (general inventory); INV – Recent Invertebrate) I.G. 11917/INV.190100, lectotype (designated herein), ♂, mantle length (ML) c. 49.8 mm, 25°11'30"N, 57°15'E (Gulf of Oman), 1,260–1,280 m (689–700 fath), 13 April 1902, bottom temperature 6.7 °C, soft green mud, Agassiz trawl, RIMS “Investigator” Stn 297 (formerly IM 1341/1). IM 350/1 paralectotype, ML unknown, radula mount (in RBINS collections), ?Andaman Sea, 496 m (271 fath), date not recorded (datNR).

Additional material examined. OUMNH.ZC.1979, ♂, ML 40.0 mm, 8570/1, ?Andaman Sea, 496 m (271 fath), datNR.

Distribution. Gulf of Oman (lectotype), ?Andaman Sea (OUMNH specimen and IM 350/1 paralectotype), to Arabian Sea and (questionably) Bay of Bengal (paralectotypes), 353–1,280 m (Fig. 1).

Description

Type material according to Massy (1916) (Table 2)

Table 2

Measurements (mm) of Teretoctopus alcocki Robson, 1932 type material in Massy (1916) (as Polypus januarii (Hoyle))

Mantle round; tissues very soft; eyes “enormous”. Colour “pinkish”. Arm pairs I or II the longest, slightly thicker than arm pair IV; arm length 3.0–4.8¹ times ventral ML². Suckers small, prominent, widely spaced, none abruptly enlarged in male. Web 25–33%³ arm length; sector E deepest, sectors A–D subequal. Mantle opens just below eye. Funnel fused to head, its aperture opening ~33% the distance between its base and the “margin” of web sector E.

Type material according to Robson (1932) (Table 3)

Table 3

Measurements (mm) and indices of type material of Teretoctopus alcocki Robson, 1932

Mantle elongate–ovoid; head narrow; eyes large (~10% total body area), prominent. Dorsal arm pair I the “largest” (?longest), 76–79% TL. Suckers small (ASIn 5.0–6.6%), none abruptly enlarged, with moderately thick-walled acetabulum. Web depth 29–34% of longest arm; sector A deepest; web formula unknown. Mantle aperture type ?B–C (sensu Robson (1925: 1334, text fig. 1)). Funnel cylindrical, rather broad, with quadripartite funnel organ: patches in specimens 1341/1 and 3348/1 subcircular, small, bead-like; patches larger and more oblong in 347/1, and larger still (represented by scars only) in 350/1.

Hectocotylised arm 53–59% of longest arm length; ligula long, narrow, 12–16% arm length, with copulatory groove constricted just above calamus, with clear laminae; sides of ligula form acute angle with calamus, enfolding its base; calamus (from figure) ~24% LL.

Lectotype (Figs 2–7, Table 4)

Figs 2–7

Teretoctopus alcocki, lectotype, ♂, ML c. 49.8 mm, RBINS I.G. 11917/INV.190100 (formerly IM 1341/1): 2 – dorsal view, entire specimen, tissue (blot) dried, MNHN post-postage, before examination; 3 – dorsal view, entire specimen, immersed in water, RBINS, pre-postage; 4, 5 – oral and lateral view of hectocotylus, respectively; 6 – quadripartite funnel organ (with outline stippled for clarity); 7 – mantle cavity and apparent viscera: a – anus; cr – crop; br – branchial heart; f – funnel; g – gill; sg – spermatophoric complex; st – stomach; t – testis; to – terminal organ; v – arm vein. Absolute measurements for characters are detailed in Table 4 and descriptive text

Table 4

Measurements (mm) of Teretoctopus alcocki Robson, 1932 lectotype (1341/1) and OUMNH non-type (8570/1) of Goodrich (1896): * – damage; + – plus more; – – no data

Mantle distorted, damaged; previously dissected; viscera incomplete. Mantle skin parchment-like, with impression of fabric (both preservation artefacts); brachial and nuchal constrictions well developed (Figs 2, 3). Points of mantle attachment to head damaged. Head narrow, (now, an artefact) almost free of brachial crown, attached to mantle by oesophagus and fragmentary mantle tissue. Funnel (now, an artefact) almost entirely free of ventral head surface; funnel aperture not constricted; funnel organ comprising four elliptical (1.4–1.5 mm length) pads (Fig. 6).

Arm widths at level where web vestiges attach to arm bases: 1R (7.0 mm), 2R (7.0 mm), 3R (6.5 mm), 4R (6.0 mm). Arms 235–289% ML, 75.4–61.3% TL, formula 1 > 2 > 4 > 3. Web obliterated. Intact arms with 140–154 suckers; suckers uncrowded, none abruptly enlarged, biserial to tips, with acetabulum wider than infundibulum; greatest diameter 4–4.2% ML, slightly smaller on arm pairs 1 and 2 than 3 and 4.

Gills distorted, almost filling pallial chamber: left, ~17 × 11 mm; right, ~25 × 11 mm (Fig. 7). Epithelium connecting ventral renal surfaces with inner mantle surfaces unapparent. Left gill with basal 4 lamellae regularly disposed, thereafter chaotic (gill distortion precludes precise differentiation of demibranchs or counting of lamellae); with ~22 lamellae in total, with terminal 2 or 3 lamellae gradually reducing in size.

Ligula long, narrow, 17.6% arm length; with ~15–17 transverse rugae (basal 9 apparent; distal 6–8 largely obliterated, inferred by remnants of scalloped margins or raised ridges) (Figs 4, 5). Copulatory groove constricted immediately distal to calamus; sides of ligula form acute angle with calamus, enfolding its base; calamus ~30.8% LL. Hectocotylised arm ~149% ML, 38.7% TL, 70.5% opposite arm length, 51.4% longest arm length; with 39 suckers.

Digestive gland missing. Remnants of alimentary canal unremarkable. Caecum (8.2 × 11 mm) pouch-like, with two hepatic ducts. Stomach (9.2 × 8.6 mm) muscular. Crop without apparent diverticulum; residual contents include fish scales; posterior oesophagus thick-walled. Posterior salivary glands broadly heart-shaped, 6.6 × 5.3 mm (13.3% ML). Buccal bulb (distance between beaks and where oesophagus enters arm crown) ~12.5 mm (~25% ML). Anal flaps absent.

Paralectotype 350/1 (Figs 8–10)

Figs 8–11

Teretoctopus alcocki, paralectotype (formerly IM 350/1), ♂, ML unknown: 8 – entire radula, dashed lines indicate where Adam (1939) may have illustrated the radula (9 – inset) from this preparation (between transverse rows 20 and 31); 10 – T. alcocki radula, redrawn from Adam (1939, text fig. 25); 11 – T. indicus radula, redrawn from Robson (1932: text fig. 46). Scale bar 1 mm (for 8); no scales were presented for figures in Robson (1932) or Adam (1939)

Radula ribbon with ~85 visible transverse rows of teeth, with 7 teeth per row, without consistent seriation (seriation repeats from rows 6 to 9 or more). Rachidian tooth damaged from rows 50–61, and worn from rows 62–85; consistently with strong basal cusp either side of medial cusp, inconsistently with single asymmetrically disposed lateral cusp on one side of medial cusp from transverse rows 1–11, and single lateral cusp either side thereafter. Laterals 1–3 and marginal blocks unremarkable.

OUMNH specimen (Figs 12–15, Table 4)

Figs 12–15

Teretoctopus alcocki, OUMNH ZC.1979, ♂, ML 40.0 mm: 12 – whole animal, dorsal; 13 – terminal portion of third right hectocotylised arm; 14 – quadripartite funnel organ (central two limbs obvious; lateral limbs less obvious); 15 – funnel organ (same image as 14, with funnel tissue digitally erased). Absolute measurements for characters are detailed in Table 4 and descriptive text

Mantle spherical, soft, smooth; nuchal constriction stronger than brachial constriction (Fig. 12); skin with numerous minute red chromatophores. Points of ventral mantle attachment to head ventrolateral to eye. Head narrower than mantle. Funnel with distal 40% free; funnel aperture not constricted.

Arm widths at points of web attachment: 1R (7.5 mm), 2R (7.1 mm), 3R (6.2 mm), 4R (6.2 mm); web extends dorsally and ventrally as narrow riblet along arms to distal 30%. Arms 287–325% ML, 65.7–74.3% TL, formula 1 = 2 > 3 = 4. Deepest web sector (C) 34.6% longest arm, 39.1% shortest non-hectocotylised arm; shallowest web sector (E) 20% longest arm, 22.6% shortest non-hectocotylised arm. Intact arms with 158–167 uncrowded suckers, none abruptly enlarged, biserial to tips, with acetabulum wider than infundibulum; greatest diameter 5.5–6.3% ML, slightly larger on arm pairs 1 and 2 than 3 and 4.

Digestive gland, gills, and reproductive system missing. Posterior salivary glands asymmetrical: broadly triangular, 7.0 × 7.0 mm (17.5% ML), almost circular, 6.1 × 6 mm (15.3% ML). Buccal bulb intact; length cannot be reliably estimated. Viscera largely destroyed; no vestige of ink sac apparent; anal flaps absent. Funnel organ quadripartite (Figs 14, 15), with central two pads (3.2 × 1.1 mm) more prominent than lateral two pads (5.1 × 1.8 mm).

Ligula long, narrow, 14.2% arm length (Fig. 13); with ~16 transverse rugae, more clearly defined basally. Copulatory groove wide, well developed; sides of ligula form acute angle with calamus, enfolding its base; calamus ~45.0% LL. Hectocotylised arm 195% ML, 44.6% TL, 67.8% opposite arm length, 60.0% longest arm length; with 55 suckers.

Remarks

While our redescription of the lectotype proffers novel information, the specimen’s poor condition and incompleteness limit what can be discerned from it; to extract its beaks, radula, and vestiges of the male reproductive and alimentary systems would destroy what limited structural integrity remains (micro-CT data for this specimen [images not presented, but available on reasonable request] revealed no characters of value). If the lectotype and OUMNH specimens are conspecific, enough information is provided to re-identify this species from specimens collected from or proximal to “syntype” (= paralectotype) localities. However, with regard to these, problems are apparent for three of them.

While Goodrich (1896) cites a collection depth of 193 fathoms and coordinates 20°17'30"N, 88°30'E for Bay of Bengal specimen 347/1, the closest station to this (14) cited by Alcock (1902) and Anon (1914) is ~40 nautical miles (nm) away (20°17'30"N, 88°51'E). A “? position” remark adjacent to coordinates for “Investigator” stn 343 in Anon (1914) also suggests uncertainty regarding coordinate accuracy for specimen 3348/1. Our plot of the “Investigator” course from stns 338–346 (Fig. 1, inset) reveals an anomalous U-turn from stn 342 off Sohar to stn 343 off Muscat (near stns 340 and 341), followed by an approximate 250 nm reversal and steam north within 24 h through the Gulf of Oman, then west and through the Strait of Hormuz to sample at stn 344. While possible, it is more plausible that station 343 (609 fath, 1,113 m) was made near but north of stn 342, and within the Gulf of Oman, because depths within the Strait of Hormuz are far shallower.

Paralectotype 350/1 was purportedly collected from the Andaman Sea in 271 fathoms (Goodrich 1896), but neither Alcock (1902) nor Anon (1914) report any Andaman Sea, Bay of Bengal, or Arabian Sea station prior to 1895 (to “Investigator” stn 182) from this depth. Because Goodrich’s account of cephalopods from (then) Calcutta Museum collections was read to the Linnean Society on 18 December 1895 (Goodrich 1896) and was based on collections received by a Professor Ray Lancaster “last year” (i.e., 1894), those specimens available to Goodrich must have been collected earlier in or prior to 1894. Stations from similar depths, or a range including 271 fath collected between 1884 and the end of 1894 include Andaman Sea stns II⁴ (272 fath), 7 (267 fath), and 13 (265 fath) (Fig. 1); Bay of Bengal stns 6 (272 fath), 129 (270 fath), 130 (258–281 fath), and 172 (200–350 fath); and Arabian Sea stn 151 (142–400 fath). Only for the herein-designated lectotype (from “Investigator” stn 297) are the coordinates unambiguous.

The metal tag and string tied around the body of the OUMNH specimen are identical to those of the herein-designated lectotype, except that the former tag has “48 B” crudely scratched into it, whereas the latter has “M ” punched into it. The OUMNH specimen was likely retained by or gifted to Goodrich and accessioned into Oxford collections on 13 January 1896 following his report on this collection (Goodrich 1896). The (new name) Indian Museum loan made out to Robson (27 November 1928) refers to only one male (M350/1) from this Andaman Sea station.

Robson (1932) described the hectocotylus of both males (350/1, 1341/1); specimens 347/1 and 3348/1 were female (Goodrich 1896, Massy 1916, respectively). Because Massy’s (1916) “end of body to eye” and Robson’s (1932) ML measurements are broadly comparable, we estimate the ML of the lectotype (1341/1) to have been ~54 mm when Robson (1932) examined it, and that of 3348/1 to have been ~60 mm. As the lectotype bears no resemblance to the specimen figured by Robson (1932: pl. III, fig. 3) for which the caption cites “type”, we surmise (given no hectocotylus is visible) that the figured specimen was the now missing 3348/1.

Although the digestive gland of the lectotype is missing, no traces of anal flaps or an ink duct remain, indicating that an ink sac was absent. We also describe the posterior salivary glands to be “vestigial” relative to ML following O’Shea (1999, table 2, p. 9). While we cannot precisely measure the buccal bulb length, nor determine its outline using micro-CT imagery, it is (at most) of moderate length (~25%) relative to the mantle according to O’Shea (1999: table 2, p. 9). Accordingly, the buccal bulb is of “normal size” for an octopus species, but the posterior salivary glands are, as described by Robson (1932), “vestigial” relative to ML for this specimen. They are “small” for the OUMNH specimen.

Two characters/states Robson (1932) considered remarkable for this species and genus were the quadripartite funnel organ, which Robson (1932) noted was almost identical for Gulf of Oman specimens 1341/1 and 3348/1, but which differed from that of the “?Andaman Sea” (350/1) and Bay of Bengal (347/1) specimens, which he maintained were also unalike, and gill size (Robson 1932: 252). Robson (1932) described the absolute size of the funnel organ as decreasing with increased ML, and he Robson (1932: 241) also commented that “I ought to remark that in none of these specimens … is the funnel-organ well preserved. It may be therefore open to question whether the very peculiar quadripartite form seen in Nos. 347, 3348 and 1341 is not purely accidental”. Although gill length is seldom reported in descriptions of octopus taxa, values reported for Muusoctopus longibrachus akambei Gleadall, Guerrero-Kommritz, Hochberg et Laptikhovsky, 2010 (24–43% ML) include and exceed those Robson (1932) reported for Teretoctopus, rendering them equally remarkable in relative length, and this character state of dubious taxonomic value (should M. longibrachus akambei be correctly attributed to this genus).

Robson (1932: 240) referred to specimen 350/1 (a paralectotype, for which only the radula remains) as so mangled that it could not be properly described, and that it was only possibly referable to T. alcocki. Because the OUMNH specimen collected with it is similar in many respects to the lectotype, and likely to be conspecific, specimen 350/1 was more probably conspecific also. Accordingly, the radula described by Adam (1939), reimaged herein (Figs 12–14), is most plausibly conspecific.

The descriptions of this species from accounts of Massy (1916) and Robson (1932) differ in details such as mantle shape (whether elongate–ovoid or round) and which web sector (A or E) is deepest. Goodrich (1896) provided no description. The OUMNH specimen (ZC.1979) is in better condition than the lectotype (RBINS I.G. 11917/INV.190100), and shares with it a quadripartite funnel organ, similar hectocotylus, and minimum and maximum arm sucker counts (lectotype 140–154; OUMNH specimen 158–167), but differs most notably (and disconcertingly) in having more (55 vs 39) suckers on the hectocotylised arm. With only two intact specimens at our disposal we cannot appraise the significance of this latter difference, although it is more than one of us (SO) would regard as “normal” variation within a species.

Teretoctopus indicus Robson, 1929

Fig. 11; Tables 5, 6

Teretoctopus indica Robson1929a: 608.

Teretoctopus indicus Robson, 1929 – Robson (1932: 249–251, pl III (fig. 2), text fig. 46).

Polypus pricei Berry, 1913 – Massy (1916; partim, specimens 8129–8131/1, 209–210)

Type locality. Arabian Sea, 24°45'N, 63°50'15"E, 995 m (544 fath), Agassiz trawl, 10 April 1906, bottom temperature 8.9 °C, “brown mud”, RIMS “Investigator” stn 366.

Etymology. Though not explicitly stated, likely from Ancient Greek ἰνδικός (indikós), meaning from Ἰνδία (India).

Description

According to Massy (1916) (Table 5)

Table 5

Measurements (mm) of Teretoctopus indicus Robson, 1929 type material in Massy (1916) (as Polypus pricei Berry)

Animal delicate, transparent. Body soft, elongate; eyes large, prominent. Arm length ~2.5 times “body length⁵”, with small, prominent suckers placed “rather far apart”, with none particularly enlarged. Web depth ~33% arm length, sector E slightly shallower, continuing little along the arms. Mantle opens “just below the eye”, funnel aperture “unusually broad”, and funnel organ (two specimens) with four pads: the median pads very small, oblong, widely spaced, and flanked by a single lateral pad (in the smallest individual, median pads 1.5 mm, separated by 3 mm). Hectocotylus with “small” terminal organ, with faint transverse rows in largest specimen, with “moderately defined” sperm canal. Body perfectly smooth, pale-buff, with minute yellow-brown chromatophores interspersed with a few larger ones. Alternating longitudinal rows of conspicuous large, light-coloured chromatophores extend along arms where skin has abraded.

According to Robson (1932) (Fig. 11, Table 6)

Table 6

Measurements (mm) and indices Teretoctopus indicus Robson, 1929 syntypes in Robson (1932)

Mantle elongate-ovoid; head distinct, very narrow; eyes large, prominent; skin smooth; mantle aperture wide (B–C). Arm lengths approximately subequal. Web very deep (WDI 35–40), formula ?A = B = C = D.E, extending little along arms, fragile. Suckers small, widely spaced, prominent. Hectocotylised arm 59–66% longest arm length; hectocotylised portion 5.5–6.3%⁶ its length; ligula flat, with few faint lamellae; calamus ~33% LL. Gills long, well developed, with 10 or 11 lamellae per demibranch; inner demibranch reduced. Radula with moderately developed cusps, with main cusp long relative to base width; first lateral with prominent cusp, wide; second lateral with long base and submarginal cusp; third lateral slender, with small base. Crop small, with diverticulum; posterior salivary glands “singularly small”; stomach “exceedingly muscular”; intestine slender, with little local dilatation. Ink sac absent.

Remarks

All syntype material of this species was sourced to collections of the Indian Museum (subsequently named the Kolkata Museum), which, until the outbreak of World War II, had been based in Calcutta. In May of 1942, these collections, along with those of the Zoological Society of India (ZSI), were moved to Varanasi into temporary accommodation at Kaiser Castle on the banks of Varuna River (Chopra 1946). A flood in 1943 (see Appendix for a more complete account) destroyed many of these wet-collection specimens. When we attempted to loan this material, it could not be located; we were also informed that it was most likely lost during this flood. To be sure, we extended our search of this material to other candidate collections, in the event this material had not been (as presumed) returned to the IM (given we had relocated specimens of T. alcocki that should also have been there).

We suspect that only one of the original three T. indicus syntypes was lost in this flood, for two of them were neither dispatched to Adam in Brussels nor returned to India, because correspondence from Robson to Adam (23 December 1932) indicates that Robson could not locate them in December 1932. Moreover, between 1927 and 1937 no record exists of any loan of cephalopods from the British Museum (Natural History) having been returned to any museum that may accidently have included these two specimens (pers. obsv.).

We note that in the second to last sentence of his description of T. indicus, Robson (1932: 251) commented (with reference to the length and form of the hectocotylus) that “the latter seem to be undeveloped in all our specimens of calmani, but even so it is very unlike that of the smallest alcocki”. From this we surmise that Robson may have intended to call T. indicus “T. calmani”, after the then Keeper of Zoology at the British Museum (Natural History), William Thomas Calman (1871–1952). However, no specimens identified as T. calmani have been located in NHM collections either.

It is possible that metal tags tied around the missing T. indicus types have been removed before being photographed for Robson’s memoir, and not replaced. If so, these two specimens may still exist, unlabelled in some collection, so it is premature to refer to either of them (or all syntype material of T. indicus, the type species of this genus) as being “lost”.

We exclude the Pearl Banks “?O. pricei (Berry)” record of Winckworth (1926) from our synonymy because insufficient information is reported for it, and we have been unable to locate the specimen. Robson appears not to have examined it either, and it has no type status. While this specimen has not been located in Colombo Museum collections (Somarathna pers. com.), it is possible that Winckworth retained it, for RBINS correspondence (28 September 1938) reveals that he donated his entire Indian Ocean cephalopod collection, including unspecified material from the Pearl Banks, to Adam. Unfortunately, no specimen labelled “?O. pricei (Berry)” of Winckworth (1926) has been located in RBINS collections to date. Nevertheless, it is premature to refer to this specimen as lost. The hectocotylised portion of the arm illustrated by Massy (1916: pl. XXIII, fig. 7) appears to be from an immature specimen.

GENERIC DIAGNOSIS

A direct comparison of T. indicus and T. alcocki is hindered by our inability to locate any T. indicus syntype, and the inconsistent nature of characters and their states, and measurements referred to and used by Massy (1916) and Robson (1932). Descriptions for these two species are also based on specimens in very different states of preservation. From what we can glean from historical texts and images, and from our own examination of T. alcocki material, little differentiates these two species. Their radulae are similar, most notably in having a rachidian tooth with a narrow base and few lateral cusps, although the teeth in T. alcocki (Adam 1939: 106, text fig. 25) are slightly shorter and stouter than those of T. indicus (Robson 1932: 250, text fig. 46). Our stacked photomicrograph of this preparation is consistent with the illustration of Adam (1939), which we believe was based on teeth closer to the base of the ribbon than its cutting edge. One noteworthy difference between species descriptions regards the crop diverticulum, which is absent in T. alcocki, but present in T. indicus. However, O’Shea (1999: 201) described a diverticulum to be present or absent within a single species referred to “Benthoctopus”, so it may be an unsuitable character upon which to differentiate species. By necessity, our generic rediagnosis is based largely on what we can discern from the more mature T. alcocki lectotype and OUMNH specimens.

Thiele (1935) reconfigured Robson’s (1932) diagnosis of Teretoctopus to “integument smooth; arms short with 2 rows of suckers and broad velar membrane; funnel organ quadripartite; radula normal; crop distinctly developed; posterior salivary glands weak”. Voss (1988) amended this to “crop with diverticulum; radula with lateral cusps on rachidian; skin smooth; funnel organ IIII” and, anomalously, with “posterior salivary glands large” (Voss 1988: 270) and contrarily “reduced” (Voss loc. cit.: 262). In accordance with O’Shea (1999: 9, table 2), for the T. alcocki lectotype, these glands are “vestigial” to “small”, ranging 13.3–17.5% ML. Our revised diagnosis for this genus incorporates this variability in the size of these glands and in crop diverticulum expression.

Molecular studies consistently reveal relationships between Muusoctopus and Vulcanoctopus to be far closer than Gleadall (2004) surmised, for the group to be paraphyletic, and for there to be no molecular support to retain these two genera as distinct (Jorgensen et al. 2010, Ibáñez et al. 2016, 2018, 2020, Sanchez et al. 2018, Strugnell et al. 2009, 2011). Unfortunately, Strugnell et al. (2009), Voight (2012), and Sanchez et al. (2018) (by exclusion of Vulcanoctopus from their phylogeny) regarded Vulcanoctopus to be a junior synonym of Muusoctopus, when the former has indisputable nomenclatural priority. Gleadall (2013: 109) acknowledged as much. We recognise Vulcanoctopus to be the senior synonym of the two, and for it to be the more appropriate of these two names to use for most taxa in Gleadall’s inkless-octopus complex.

While Robson (1932) was uncertain if the quadripartite funnel organ of Teretoctopus was real or an artefact, it is well preserved in the OUMNH specimen. This organ in the T. alcocki lectotype and Ameloctopus (Norman 1992: fig 1e–g) are for all intents and purposes identical, and they are otherwise very similar in the OUMNH specimen to that of V. normani (Massy, 1907) (new comb.) (Allcock et al. 2006: 377–379, fig. 3C), and to a lesser extent, Bathypolypus bairdii Verrill, 1873 and Bathypolypus pugniger Muus, 2002 (Muus 2002: 208, fig. 18). The other character state that Robson (1932) considered remarkable for Teretoctopus, relative gill length, is also not unique to this genus. Therefore, because no single character or state described for Teretoctopus is unique to this genus, the better-preserved OUMNH T. alcocki looks very much like other species formerly attributed to Muusoctopus or Benthoctopus, and Teretoctopus has nomenclatural priority over Vulcanoctopus, Teretoctopus may prove to be the most appropriate name to apply to species in this complex of three genera. While we lean towards this course of action, we stress that no such synonymy should be made until molecular data for T. indicus or T. alcocki from specimens collected from or proximal to the type locality (consistent with redescriptions herein) is procured. For this reason, we describe these specimens in as much detail as possible to facilitate reidentification of this taxon.