New insights into the morphology, reproduction and distribution of the large-tuberculate octopus Graneledone macrotyla from the Patagonian slope

The new information reported in this paper is based on 11 specimens of the large-tuberculate octopus Graneledone macrotyla. These specimens were caught in bottom trawl surveys ATLANTIS 2009 and 2010 carried out on the Patagonian slope off the Argentinean Economic Exclusive Zone between 24 February and 1 April 2009 and from 9 March to 5 April 2010 respectively. A new diagnosis and a complete description of the species are provided. This is the first time that stylets, beaks and spermatophores are described. This is also the first time in which mature females have been studied and the female genitalia described. Like other eledonid octopods, G. macrotyla does not have spermathecae in the oviducal glands. The presence of fertilized eggs inside the ovary suggests that fertilization takes place within the ovary. The simultaneous occurrence of oocyte cohorts at different oogenic stages suggests that the species is a multiple spawner. G. macrotyla inhabits shallower waters on the Patagonian slope (475-921 m) than in the subantartic area (1647-2044 m). From a biogeographical point of view, our data show that G. macrotyla inhabits the plume of cold subantarctic waters, which is pushed far north into the southwestern Atlantic by the Falkland (Malvinas) Current.


INTRODUCTION
The large-tuberculate octopus Graneledone macrotyla Voss, 1976 was first described on the basis of a single specimen: a female, mantle length 34.5 mm, caught by the USNS ELTANIN at Station 1592, 54°43'S, 55°30'W in 1647-2044 m depth with a 10-foot Blake trawl on 14 March 1966 (Voss 1976). Furuya and Hochberg (2001) described a new species of parasite, Dicyemennea dorycephalum, from the holotype. Kubodera and Okutani (1994) provided a supplementary description and some remarks on this octopus species, including details of the male reproductive organs and new distribution records, based on two specimens caught south of 45ºS along the southeastern Patagonian slope.
In spite of this additional information, the description of this species is still incomplete, mainly because the three specimens used were small (ML 34.5-67.0 mm) and did not include any mature females.
The main aims of the present paper are to improve and complete the description of G. macrotyla and offer new insights into its geographic distribution and reproductive strategy.

MATERIALS AND METHODS
Eleven specimens of G. macrotyla were caught during two multidisciplinary research cruises conducted by the Instituto Español de Oceanografía (IEO) to assess the biomass of the main commercial stocks on the high seas of the southwest Atlantic. The bottom trawl surveys ATLANTIS 2009 and 2010 were carried out between 24 February and 1 April 2009 (Guerra et al. 2011) and from 9 March to 5 April 2010 respectively, on board the R/V Miguel Oliver (Table 1).
Catches were sorted immediately on board after capture. Cephalopods were stored in labelled plastic bags and frozen at -20ºC. Specimens were subsequently transported from the R/V Miguel Oliver to the Instituto de Investigaciones Marinas (IIM, CSIC) in Vigo, Spain, for detailed examination. The material used in this work, as well as the entire collection of cephalopods collected in these two research cruises, are deposited in the Museo do Mar de Galicia (MDMG) at Vigo, Spain (www.museodomar.com/es).
After thawing at room temperature the specimens were identified and the sex was determined. Total length (TL), dorsal mantle length (ML) and wet weight were measured. The 11 specimens of G. macrotyla were then preserved in ethanol 70%. All measurements were repeated on these preserved specimens. We used the morphometry defined by Roper and Voss (1983).
The histological study of the oviducal gland and oviducts was conducted on specimen MDMG 68B2011, whose gland had a maximum diameter of 27 mm. These structures were fixed in a solution of 10% buffered formaldehyde at a formaldehyde/tissue ratio approximately equal to 50. After fixation the tissues were thoroughly rinsed with water. Each sample was then dehydrated in ethanol in increasing concentrations from 70 to 100%, and embedded individually in a paraffin block (58ºC) according to standard histological techniques (Gabe 1968). Longitudinal and transverse sections were cut (6-8 μm thick) at different levels of each structure and stained with HARRY's hematoxylin-eosin. Sections were photographed and examined using an image analysis system (NIS-Elements 3.0).
The radula preparation procedure was as follows: tissues around the radular ribbon were digested with trypsine 1% during one day, and then the radula was mounted in glycerine, dehydrated in ethanol 96%, cleared with xilol, stained with Orange G to 0.1% during 1 hour, and mounted in synthetic resin DPX. Due to teeth variations, especially in the rachidian tooth, the most common teeth pattern has been drawn. The mounted radula was photographed using a binocular microscope with a magnification of 60x coupled with an image analysis system (NIS-Elements 3.0). Radular nomenclature follows Nixon (1995).
Given the lack of genetic data for this species, tissue samples have been submitted to the International Barcode of Life (iBOL) in order to amplify the barcoding cytochrome oxidase subunit I (COI) gene. Voss 1976 (Figs 1-10, Tables 2-3) Diagnosis. Medium size species, mature males from 84 mm ML and mature females from 110 mm ML, with distinctive sculpture. Dorsal surface of mantle, head and basal part of the arms covered by distinct complex papillose warts; 10-12 wart clusters across the dorsal mantle. Body muscular. Arms typically 1.8-2.5 times ML. Two tongue-shaped and tuberculate supraocular cirri on each eye, the posterior larger (10-14% ML) than the anterior one.

Graneledone macrotyla
Description. Medium-sized octopodid; ML up to at least 110 mm for females; TL to at least 366 mm  .0). Two tongue-shaped and tuberculate supraocular cirri on each eye, posterior cirrus larger (10-14% ML) than the anterior cirrus (5-6% ML) (Fig. 3). Male and two females with posterior supraocular cirrus only, lengths of which varied from 9.9 to 12.0% ML. Slight constriction between mantle and arm bases.
Digestive tract illustrated in Figure 3d of Voss (1976). The buccal mass large. Anterior salivary glands small. Posterior salivary glands lanceolated and relatively small (16-17% ML). Oesophagus stout and slightly dilated along much of its length, without diverticulum and apparently not forming a true crop. Dilated portion of oesophagus leads into a stronglydifferentiated two-part stomach and a distinct large and little spiral caecum. The intestine is large, slightly dilated, and bent almost double upon itself. The anal pore is round. Anal flaps absent. No trace of ink sac.   Upper and lower mandibles or beaks well developed, not markedly different from other octopodids (Fig. 6). Radula shown in Figure 7. Well developed testis, first section of the duct system or proximal vas deferens narrow and convoluted, long spermatophoric gland, two strong coils proximally, the more distal one almost three times larger than the proximal one, long and thick second spermatophoric or accessory gland, thin Needham's or spermatophore storage sac, thin and poorly defined medial and distal spermatophoric ducts, terminal organ or penis curled, moderately long (21.2% of the ML) and with a large diverticulum, sac-like and weakly coiled. Spermatophores very long (SpLI 110.4) and apparently produced in low numbers (2 in Needham's Sac in mature male MDMG742011, Table 2). Oral cap of the spermatophore hooked and very distinct ( Fig. 8A and 8B).
Female genitalia illustrated in Fig. 9. Mature ovary oval-shaped. Proximal oviducts robust and short (6-9% ML); distal oviducts relatively long (52-69% ML) and also robust (Fig 9A and 9B). Oviducal glands large; diameter varying from 15.4% ML to 23.3% ML in submature and mature animals respectively. Oviducts change colour with maturity from white-beige to uniformly dark grey-blueish when mature; no radiating chambers at any maturation stage.  Eggs large, mature oocytes 20.5-25.5 mm long (EgLI 18.6-21.1), wide (EgWI 8.7-10.7), and produced in relatively low numbers (263 in specimen MDMG68C2011, Table 3). All five ovaries contained oocytes. Oocytes in four different developmental stages. Immature ovaries (specimens MDMG 94A2011 and 94B2011) contained immature oocytes, which have a smooth surface with some brown spots and different sizes (maximum length 1.0-3.0 mm). Maturat-ing ovaries (specimens MDMG 552011 and 68C2011) contained immature, maturing (3.1-12.8 mm) and fully mature oocytes of different sizes (18.8-25.5 mm). Mature ovaries contained fully mature oocytes, and few (3-4) fertilized eggs (18.8-20.5 mm). All oocytes at these stages showed a rigid and translucent outer cage with longitudinal striae, narrow micropyle at the distal pole, and long peduncles (Fig. 9C); they were attached to the ovary like clusters of grapes. The fertilized eggs Table 3. -Graneledone macrotyla. Indices of the female holotype (USNM727678), the two specimens (67838 and 67839) in Kubodera and Okutani (1994)   were oval in shape (similar to a hen's egg), external capsule hardened, smooth and beige in colour, no micropyle, and very short peduncles (Fig. 9D). Oviducal gland in the mid-portion of each oviduct, formed like a huge concentric peripherical gland around the oviduct and separated by a thin sheet of connective tissue; spermathecae absent. Peripherical gland is formed by groups of concentric cells with basal nuclei and a central lumen; in females close to maturity the cytoplasm is densely packed with reddish grains (Fig. 10).
Spermatozoa not seen associated with oviducts. No spermatangia or free sperm within ovaries or attached to egg filaments. Body sculpture distinctive. Low, thin, peripherical keel around mantle. Dorsal surface of mantle above keel, head and basal part of arms covered by distinct complex papillose warts 3-10 mm in diameter, consisting of a pale raised area bearing small simple 4-10 tubercles (papillae) generally surrounding one erect, central, sharp tubercle. Ten to twelve wart clusters across dorsal mantle. Warts extend onto web and bases of arms for approximately 15% of arm length, except on arms 3 and 4 where warts are absent. Row of papillose warts around each eye. No papillose warts on ventral surface of mantle and head. Base colour on dorsal surface of animal ranges from brownish red to purple. Colour is paler on oral surface. Raised pale areas creamy brown.
Distribution. Figure 11 shows the records of G. macrotyla described in the present paper, the two records described by Kubodera and Okutani (1994), and the holotype (Voss 1976). Our records extend the known distribution from 45°44'S to 45°08.02'S and the known depth range from 475 to 2044 m, where the holotype was caught (Voss 1976).

DISCUSSION
Within the genus Graneledone Joubin 1918, G. macrotyla has an unusual wart structure that makes it very distinct from the other seven species of the genus (Guerra et al. 2000;Allcock et al. 2003). The diagnostic characters of the specimens described here agree with those indicated by Voss (1976) and Kubodera and Okutani (1994). Morphometric measurements and counts are generally within the ranges given by these authors. The discrepancies observed could be due to conservation. This is the first time that stylets, beaks and spermatophores are described, and also the first time that mature females were found. The asymmetry observed in the stylets might be due to body distortion caused by freezing. The radula structure of G. macrotyla was variable (homodont and heterodont); the rachidean tooth showed one, two or three cusps depending on its location in the radular ribbon. The use of the radula as a diagnostic character in cephalopods to the generic and specific levels is still controversial, mainly because the radula shows a high individual degree of variation that is evident with both light and scanning electron microscope (e.g. Adam 1941 in different octopod species; Allcock et al. 2003 within the genus Graneledone). Nixon (1998) observed this problem and suggested a number of morphometric studies to ascertain the use of the radula as a diagnostic character; however, they have not yet been carried out. For the above reasons, we decided not to include the shape of the radula as a diagnostic character in G. macrotyla.
The absence of spermathecae in the oviducal glands was observed in four species of Eledone (Perez et al. 1990). This is the first time that this character has been shown in a Graneledone species. Voss and Pearcy (1990) reported that the distal oviduct of Graneledone pacifica contained sperm. However, given that no spermatangia or free sperm were observed within the ovaries of this species, the presence of fertilized eggs inside the ovary suggests that fertilization takes place within the ovary, as observed in Eledone massyae, E. gaucha and E. cirrhosa (Perez et al. 1990). This fact also suggests that, as occurs in the two South American eledoninids, G. macrotyla males become mature early and remain sexually active for a greater part of their life than females. It is also possible that females of this species can copulate a considerable number of times before maturation and store sperm until fertilization  Kubodera and Okutani (1994); circles, present records.
shortly before spawning. It is still unknown where the sperm is attached to the ovaries. The existence of oocyte cohorts at different oogenic stages suggests a continuous spawning reproductive strategy, in which ovulation is characterized by continuous asynchronous production of ova in the ovary after spawning has commenced. In this monocyclic spawning pattern, adults spawn continuously during their relatively extended life spans, egg-laying occurs in separate batches, and somatic growth takes place during spawning events (Rocha et al. 2001). This strategy is also followed by several cirroctopods and some species of octopods inhabiting deep cold waters (e.g. Bathypolypus arcticus and B. piscatorium) as well as Graneledone pacifica (Bello 2006). A continuous spawning strategy is common in extremely stable environments, and could be an adaptive response to 'safe' environments where predation pressure and competition for resources may be very low. In such environments, spawning can be slow and continuous, with a limited number of eggs per clutch (Rocha et al. 2001).
It is noteworthy that there were numerous oocytes with translucent, longitudinally striated outer cases and long peduncles, but with outer widely opened cases in the distal pole and with a flaccid consistency (Fig. 9D). The high salt content of marine invertebrate cells leads to extensive cellular damage during freezing (Dixon et al. 2002), which induces artificial cleavage of apoptosis-related proteins in human bone narrow (Schmidt-Mende et al. 2000). Atresia has been observed in several species of cephalopods (Melo and Sauer 1998;Boyle and Chevis 1991;Laptikhovsky 2001;Laptikhovsky and Arkhipkin 2001), and massive atresia is more common than we might think, particularly in fish (Rideout and Tomkiewicz 2011). Further analyses will be needed to elucidate whether the appearance of these oocytes is a matter of apoptosis or massive atresia. However, our opinion is that it is most likely an artefact produced by cell damage due to preservation, since the animals were frozen for 12 and 5 months for ATLANTIS 2009 and 2010 respectively G. macrotyla is a rare, bathybenthic species known only from a very few specimens captured in the west Scotia Basin (Voss 1976;Kubodera and Okutani 1994). The species is currently recorded from depths of 850-2044 m, but four of our specimens (MDMG742011, 1082011, 472011 and 942011) indicate that the species also lives in shallower waters. From a biogeographical point of view our data show that G. macrotyla inhabits the plume of cold subantarctic waters, which is pushed far into the South Atlantic by the Falkland (Malvinas) Current. This concurs with Strugnell et al. (2008) who showed that the deep-sea lineage of these octopods has its evolutionary origins in Antarctica. We suggest that future sampling should be carried out on both sides of this plume, in regions where there is no Antarctic water, to determine whether this cold-water adapted species is present.