A new species of pygmy Paroctopus (Cephalopoda: Octopodidae): the smallest southwestern Atlantic octopod, found in sea debris

The new species, Paroctopus cthulu sp. nov. Leite, Haimovici, Lima and Lima, was recorded from very shallow coastal waters on sandy/muddy and shelter-poor bottoms with natural and human-origin debris. It is a small octopus, adults are less than 35 mm mantle length (ML) and weigh around 15 g. It has short to medium sized arms, enlarged suckers on the arms of both males and females, large posterior salivary glands (25 %ML), a relatively large beak (9 % ML) and medium to large mature eggs (3.5 to > 9 mm). The characteristics of hatchlings of two brooding females, some of their anatomical features, and in-situ observations of their behaviour are a clue to the life history of it and closely related pygmy octopuses. The Bayesian phylogenetic analysis showed that Paroctopus cthulu sp.nov. specimens grouped in a well-supported clade of Paroctopus species, separate from P.joubini and P. cf mercatoris from the Northwestern Atlantic . The description of this new species, living in a novel habitat of human debris in shallow water off Brazil, offered an opportunity not only to evaluate the relationship among the small octopuses of the western Atlantic, Caribbean and eastern Pacic, but also their adaptation to the Anthropocene period. hatchlings a microscope their morphology and chromatophore The were obtained for the descriptions of eggs and hatchlings: egg index = egg length × 100/adult ML), hatchling size index = hatchling ML ×100/ adult hatchling AL index = of hatchling x 100/ML hatchlings, to Boletzky Boletzky et and et

found on the sea bottom, including metallic cans, glass bottles and plastic objects such as snorkeling mouthpieces, among others. No specimen was found inside empty shells, although we looked for them, also 9 divers. Two females with eggs were found, one spawned inside a snorkeling mouthpiece and the other in a metal beer can.
Unless stated otherwise, all measurements are in mm and weights in g. Small structures such as the ligula, calamus, spermatophores and eggs were measured with an ocular micrometer in a binocular microscope.
The description was based on adult males with fully formed spermatophores, ligula and enlarged suckers, adult or spawned females with developing oocytes or spawned eggs and some sub adult specimens in which the sex could not be determined. All the specimens evaluated for this paper were deposited in the mollusk collections of the MORG and MCPUCRS.
Body patterns and behavioral postures were photographed and lmed during dives or inside an aquarium. The chromatic, skin texture and body patterns components were described following Mather (1972) and Hanlon (1988).
Eggs, embryos and hatchlings were described after collection of a brooding female inside an aluminum can in Rio de Janeiro (Praia Vermelha Beach). Live eggs and hatchlings were lmed, xed in alcohol 70% and then sent to the Cephalopod Early Life Stages Laboratory at the University of Parana, PR, Brazil.
The eggs and hatchlings were analyzed and measured under a stereo microscope and their morphology and chromatophore patterns described. The following indices were obtained for the descriptions of eggs and hatchlings: egg index = egg length × 100/adult ML), hatchling size index = hatchling ML ×100/ adult ML), hatchling AL index = AL of hatchling x 100/ML hatchlings, according to Boletzky (1974), Boletzky et al. (2002) and Hochberg et al. (1992).

Molecular data and analyses
Small pieces of mantle or arms of three specimens were preserved in 99 % ethanol from which genomic DNA was extracted using the GF 1 Nucleic Acid Extraction kit (Vivantis, Malaysia) according to the manufacturer's instructions. Sequences of 33 species were also retrieved from GenBank (Table 1).
Fragments of the mitochondrial cytochrome oxidase subunit I (COI) gene were ampli ed by using the universal primers LCO1490 and HCO2198 (Folmer et al.1994). The PCR ampli cation reactions were conducted in a nal volume of 25 µL containing 1 µL forward primer, 1 µL reverse primer (10 mM), 12.5 µL Taq DNA Polymerase Master Mix (Ampliqon A/S) or MyTaq RedMix (Bioline), 8.5 µL H 2 O and 2 µL DNA (20 -40 ng/μl). Ampli cation PCR cycle parameters were: 3 min at 95°C for denaturation, followed by 35 cycles of 1 min at 94°C, 1 min at 45°C for annealing, 1.5 min at 72°C for extension, and a final extension step of 4 min at 72°C. The PCR products were puri ed and sequenced by Macrogen Inc, Seoul, Korea. Electropherograms were edited with Geneious 9.0.2 (Kearse et al.2012) and sequences were aligned by ClustalW using Mega 6 (Tamura et al.2013). The substitution model GTR+G was chosen using the software jModeltest (Posada 2008).
Bayesian phylogenetic inference was carried out in BEAST 1.8.4 (Drummond et al.2012). An uncorrelated lognormal relaxed clock model was used. Monte Carlo Markov Chain (MCMC) run were performed for 1×10 8 generations, sampling one tree each 1×10 4 runs. The convergence of MCMC runs, effective sample size, and the correct 'burn in' for the analysis were assessed using Tracer v1.6 (Rambaut et al.2014). A consensus tree accessing the posterior probability values of each clade was generated using TreeAnnotator 1.8.3 ) and displayed using FigTree 1.4.3. Pairwise genetic distances of Paroctopus species were calculated using K2P distance in MEGA 6 (Tamura et al. 2013) (see Table 1 of octopod specimens used to construct the Bayesian phylogenetic tree in this study).
Genus ampli ed diagnostic features: Small bodied; mantle short and wide, with large eggs. Stylets present, non mineralized. Arms short medium sized, stocky, 3 to 4 times mantle length. One to three enlarged suckers on the arms of males. Copulatory organ (ligula) medium size; calamus short. Gills with 5 8 lamellae per outer demibranch. Oviducal glands without braiding chambers. Spawned eggs with 4.2 9 mm total size), attached singly with stalks to shells, hard bottom or objects in small clusters. In living specimens body uniformly colored with little pattern variability; patch and groove system absent; white spots absent from dorsal mantle and large arm base; frontal white spot complex presents but faint. Skin without large primary papillae.  Etymology: The name cthulu is a term with a dual allusion. First, it is an irony due to the small size of the new octopus species, compared to the giant ctional entity 'Cthulhu', created by Lovecraft (1928) and described as resembling an octopus, a dragon and a human caricature. Second, it refers to the proposal of Donna Haraway of the Chthulucene as a diverse Earth wide tentacular power of symbiosis. Chthulucene proposes a holistic and biocentric coexistence that will integrate and transform the far less optimistic view of the Anthropocene (Haraway 2015). Most P. cthulu sp. nov. specimens were found in metal and plastic debris, suggesting that the species is utilizing the garbage in oceans, as an alert to this increasing global threat to the marine biota.
Diagnosis: adults small sized (ML 14 to 33 mm), mantle and head wide with large and prominent eyes. One to three cirri over the eye and one above. Shallow web and thick arms subequal in length, three and a half times the ML. One to three enlarged suckers located on the 5th or 6th row in some or all arms of all adult males and few adult females. Third arm of the males hectocotylized with a moderate calamus and small ligula, of the length of the opposite arm (Fig.  3b). Gills with 5 6 lamellae per hemibranch, usually 6. Adult females with medium to large oocytes in the ovaries (4.7 to 9 mm length). Spawned eggs attached with stalks to objects singly in small clusters. Hatchlings have 5.0 5.2 mm total length and 2.52 mm ML, arms with 14 16 large suckers. Body color of adult animals in the environment varies from yellow to reddish brown. Ventral surface of mantle, head and web with small well spaced papillae, dorsal mantle and head with larger papillae. Brownish red smooth dorsal mantle surface in preserved specimens.
Description: The following description is based on the holotype (male) and 8 mature paratypes (3 females and 5 males) and 3 subadults morphotypes (1 female, 2 males) (Tables 3, 4 and 5). Small size adults (ML up to 32.7 mm; 13 g total in preserved specimens). Broad mantle with a muscular wall (MWI 55 80) (min max) (Fig. 4a); posterior mantle with a bi lobed shape ventrally (Fig. 4b); head wide (47 72); funnel tubular (FLI 19 54), with almost half of it free (FFLI 39 63). U U shaped funnel organ with similar size (Fig. 3c). Normal arms similar in length (ALI 270 450, in mean 355), no consistent arms formulae; arms width (AWI 11 20). Total number of suckers on the normal arms 112 to 174, mean 15. Normal sucker diameter moderate (SDnI 8 13 in mean 0.105). One to three conspicuous enlarged suckers (SDeI 10 23, mean 0.15S), present in two to four arms of all examined adult males and one adult female, located at 5th or 6th row 5) ( Fig. 3c and 3d). Third right arm of males hectocotylized, moderately shorter than the opposite arm (OAI 45 73), bearing 86 94 suckers; hectocotylus with a well de ned spermatophoric groove, relatively short hectocotylus (HLI 5 -10), a moderately long calamus (CLI 20 42) and small ligula (LLI 3.7 7.3) with clear transverse trabeculae. Web shallow (WDI 20 72), with no consistent web formulae differences among the evaluated individuals, however the sectors A and E usually are shallower. Hemibranchs have stout gills, with 5 6 lamellae per hemibranch, usually 6. Stylets medium sized 5.4 6.7 mm (before preservation), with the posterior part longer and thinner Digestive tract: a dissected adult female (32.0 mm ML MORG 52768) presented a typical Octopus digestive tract (Fig. 5a), with few particularities. Large buccal mass (6 mm; 19 % of ML); pair of attened, medium sized anterior salivary glands (1.8 mm, 5.30 %ML), and large posterior salivary glands triangular (8 mm; 25% of ML) joined by ducts to the buccal mass. Narrow oesophagus followed by crop diverticulum and a wide stomach. Spiral caecum connected by two ducts to large digestive gland (12 mm, 37% of ML); ink sac embedded in digestive gland surface. Intestine relatively short and curved with a loop, ending in muscular rectum with anal aps. Beak, relatively large if compared with the species size, 1.7 mm of upper hood length (9% of ML); prominent rostrum and sharp rostral tip (upper rostral length 0.5mm), with narrow wings (Fig.5b c beak). Rachidian tooth on a half radula; two lateral teeth, one marginal tooth, one marginal plate, one lateral cusp on each side of rachidian tooth with a symmetric seriation, the cusp goes from the base to the middle of the tooth every one or two teeth (A 1-2); cusp on outer margin of rst lateral tooth; second lateral tooth triangular, almost symmetrical; marginal tooth thin, curved; marginal plate small, at (Fig. 4d).
Female reproductive system. The mature female (32. mm ML) has a very large round ovary (22 mm wide), occupying almost the whole posterior portion of the mantle, two short proximal oviducts (5.7 mm), two small spherical oviducal glands (3.7 mm), reddish orange in color, and a medium size distal oviducts (10.7 mm) (Fig. 6a). We counted a total of 30 oocytes inside the ovary. The three mature females (21.3 to 32.0 mm) showed oocytes varying from medium to large size compared to the mantle length (from 4.7 to 9 mm) (Fig. 6b).
Male reproductive system: The holotype mature male (18.3 mm ML) had a testis of 4.5 mm length, which is relatively large in the system; vas deferens narrow, with turns and wrapped in a membranous sac. Vas deferens opening into a spermatophore gland, long and curved accessory gland, both opening in an atrium linked to a long and wide Needham's sac, with almost the same size as the testis; small terminal organ tubular (PLI 1.8 6) diverticulum not clearly differentiated from the terminal organ (Fig. 6c). Spermatophores medium-sized (SpLI 39.9 43.7), narrow, without swelling (SpWI 3.5); medium sized sperm masses (SpRLI 52.1), 19 20 turns on the spermatic mass (Fig. 6d). The maximum number of spermatophores counted in the Needham's sac was 13.
Brooded eggs, embryos and hatchlings: a female with more than 30 eggs were found attached individually to the snorkel mouthpiece (Fig. 7a). Another female with spawning were found with 124 eggs were found attached individually to the aluminum can by a thin chorion stalk (2.57± 0.18 mm, n=25), along with empty chorions, as many individuals had hatched.. These eggs were medium sized, elongated to pear shape with a mean length of 4.61±0.35 mm and largest mean width of 2.3±0.14 mm (n=25). (Fig. 7b) The egg index was 23. They have a transparent chorion and were devoid of other capsules. The eggs were at different developmental stages, all of them before the second embryo inversion (stages XII.1 -XIX.1, Deryckere et al. (2020)), indicating that spawning took place over several days. Late-stage embryos (stages XVIII XIX.1) had large darkish eyes with a whitish retina and a mean eye diameter of 0.42 ± 0.1 mm (Fig. 7b). All the eight arms were well developed and similar sizes, having from 10 12 suckers distributed in a single row from the buccal mass up to the web close to the base of the arms and in a zigzag double row along the length of the arms. Around the buccal mass, there was a single to double sucker ring formed by the single row of suckers up to the base of each arm. The funnel was long, wide and conspicuous, reaching the base of the ventral arms.
Embryo chromatophores: the preserved embryo has a large number of dark chromatophores. Dorsal view: On the arms there are from 12 14 chromatophores, two at the base in a single row and the others in a zig zag row. On the head, there are about 20 24, both large extra-tegumental and small tegumental chromatophores interpolated and sometimes superimposed. On the mantle, there are from 18 to 22 extra-tegumental chromatophores distributed in the central area. Ventral view: On the head, there are four, two very large extra-tegumental chromatophores on the lateral sides of the funnel; over the mantle there are from 61 to 72 brownish large chromatophores that seems to be distributed in 8 10 horizontal rows, but when expanded cover the entire surface of the mantle. When the chromatophores are all expanded, the embryo has a dark coloration (Fig. 7c).
Hatchlings: Hatchlings have a total length of 5.0 5.4 mm ( Fig. 7d and Fig. 8a−c). The mantle edge is clearly visible at the base of the head, indicating a ML of 2.52 mm, or 3.3 mm ML if the middle point between the eyes is considered instead. The mantle is roundish with a width of 1.9 mm. The head is wider (100% ML) than long (50% ML), with approximate concave shape. The eyes are large and prominent (28% ML) and anteriorly inserted ( Fig. 8a and b). The arms are long relative to the mantle (80% ML) and robust, with 14 to 16 suckers each. A conspicuous web is present at the base of all arms. The suckers are arranged in a biserial zigzag row, as in the embryo, and the size of the suckers decreases from the base towards the tips of the arms. The base of the arms occupies a narrow area in relation to the width of the head, leaving an empty space between the head and the arms, which gives the appearance of a short arm crown stalk ( Fig. 8a and b). The body of the whole animal is covered by an unpigmented transparent skin lm, with the exception of the aboral surface of the arms, which is lled with suckers. This skin lm seems to be continuous over the whole dorsal side of the body and the only apertures are found ventrally, at the mantle edge and at the funnel ori ce. This is particularly evident in a lateral view (Fig. 8c), where it is possible to notice that the skin lm forms a large and continuous dorsal cavity from the mantle to the tips of the arms and also ventrally from the arms tips to the funnel.
Thus, the hatchling has a large continuum dorsal cavity and both ventral mantle and head cavities, the internal organs occupy a relatively small area inside the mantle. Particularly on the mantle the skin is densely covered by Kölliker's organs, which gives a rough appearance. The skin lm is conspicuous around the arms and head giving the whole animal a transparent to whitish color less silhouette when the chromatophores are contracted. When all the chromatophores are expanded, however, the hatchling acquires a very dark pigmentation as described below.
Hatchlings chromatophore pattern: on dorsal view, the chromatophore pattern on the arms seems to follow that of the suckers. There are from 10 14 chromatophores on each arm, one to two large ones at the base (in the arm crown stalk) and the others distributed in a zigzag row from the base toward the tips of each arm. On the head, there are 24 chromatophores, 17 darkish brown and seven yellow. In the anterior region, close to the base of the arms, there are four distributed in a rhombus shape, three yellow and a dark one; two larger dark ones between the eyes; six dark ones forming a row on midhead; four large dark ones at the base of the head and four close to the eyes (two dark ones interpolated by two yellow ones). On the mantle only dark chromatophores were observed. There is a double row of chromatophores around the whole mantle edge with about 12 16 chromatophores each and the same patterns is seen on the posterior mantle, where each row has about 12 14 chromatophores. Many small other chromatophores are found scattered over the whole mantle without a particular pattern. There are from 8 to 11 extra-tegumental chromatophores covering the visceral mass arranged in an oval shape (Fig. 8a). Ventral view: The distribution of chromatophores on the arms is the same as described for the dorsal view. The head has eight chromatophores, two yellow ones on the sides of each eye, two dark ones between the eyes and two very large dark ones on the sides of the funnel. Over the funnel there are six. The whole mantle is densely covered by approximately 70-80 chromatophores, which are distributed in 8 10 irregular rows. When all the chromatophores are expanded the mantle is entirely dark (Fig. 8b). Lateral view: Over the head it is possible to notice two other chromatophores underneath the eyes, a small dark one and a large yellow chromatophore close to the mantle edge. On the ventral mantle, the single row of large chromatophores around the mantle edge is clearly seen (Fig. 8c).
Adult Body Pattern: xed specimens (without previous freezing) had smooth skin on the dorsal surface (Fig. 4). Color in xed specimens varied from light brown to light reddish, darker around the eyes on the dorsal surface and clearer cream color on the ventral, with fewer chromatophores. The dorsal mantle with small papillae around the eyes was visible only in few specimens (Fig 4a and b).
We observed ve main body patterns in living animals ( Fig. 9): 1. uniform reddish with dark eyes (a); 2. Uniform dark brownish (b); 3. uniform light brown with white dots (c); 4. mottle with yellow blotch and white spots (d); 5. Brown with white stripes and blotches on arms and mantle (e). We only observed a patch and groove trellis arrangement on the dorsal mantle during the patterns Uniform light brown with white dots and Brown and white stripes. Three chromatophore colors were identi ed (Red, Brown, and Yellow). The brown and red colors could be widespread throughout the whole body (Fig. 9a, 9b), while white (no chromatophores expanded) and yellow were concentrated in localized areas: yellow appeared as blotches on the dorsal mantle, head and proximal arms areas (Fig. 9c); while small white dots were spread across the whole body (9d), and as two frontal white circles. The skin texture was characterized by three primary papillae around the eyes and smaller ones spread throughout mantle, head and proximo dorsally on the rst arms.
Distinguishing postures: We observed three stereotyped postures: sitting with curled arms pointed, (9a; 9c); sitting with eyes raised (9e) and, the rst pair of dorsal arms up showing the larger suckers on the aboral surface of the arms, (Fig. 9f)

Habitat and in vivo observations
There is no information on the habitat of the four specimens deposited in the MORG and MCPUCRS collections. Those collected in 2014 and 2015 in Rio de Janeiro were found during the daytime at 0.5 to 5 m depth, on sandy or muddy bottoms near rocky shores (Fig. 10), inside metallic cans, plastic objects or glass bottles. The specimens came out of the debris as soon as they were taken out of the water. No specimen was collected from shells.
The debris occupied varied in preservation, some of the cans were fragmented and rusty ( Fig. 10a and, b), others were intact with some biological encrustation, and few were intact and well preserved. Two spawned females were observed in vivo in their habitat. One, among the collections in February 2015 at Ilha Grande, was found inside a plastic snorkel mouthpiece with eggs attached singly in small clusters (Fig. 7a) at 6 m depth, and a water temperature around 22 ˚C, during summer time. The second female was followed for three weeks at Praia da Urca, Rio de Janeiro. She was found inside an aluminum can, at a shallow depth (2 m), and temperatures around 25 ˚C, from February to March.

Discussion
Our study revealed a new species of the genus Parocotopus, the rst pygmy octopus described for the Southwestern Atlantic. Both morphological and molecular analyses corroborate the great divergence of P. cthulu sp. nov. from the Northern Atlantic complex of pygmy octopuses, whose taxonomy is still not solved (Jereb et al.2014).

Molecular phylogeny and geographic distribution
The genetic distances between P. cthulu and P. joubini/P. cf mercatoris are large enough (around 9%) to claim that the linage from Brazil is a different species of pygmy octopus from those in Florida, USA. Additionally, the COI sequences from P. joubini and P. cf mercatoris are identical, which means either a misidenti cation problem or the species are synonymous. Misidenti cations in other Atlantic octopod species were found previously, and coupling morphological, molecular, and ecological data, have been useful to address taxonomic uncertainties (Lima et al., 2017).
The Western Atlantic pygmy octopuses probably shared a common ancestor before the uplift of the Isthmus of Panama, which is evidenced by their close relationship with P. digueti and Octopus alecto from the East Paci c (Lima et al.2020). Paroctopus cthulu may have arrived in the Southwest Atlantic via shallow-water of the continental shelf linking South and Central America, before the effects of Amazon river discharge in the Atlantic Ocean around 10 million years ago (Mya) (Hoorn, 1994), which became a low salinity barrier for many marine species (Muss et al. 2001, Rocha 2003, Gleadall 2013. This event coincided with the split between P. cthulu and P. joubini (mean 9.4 Mya) according to Lima et al. (2020). The Brazilian pygmy octopus probably settled in the Southeast and South of Brazil due to its preference for subtropical waters. Until now, we only have recorded it from Espirito Santo (20º19'09' S and 40°20'50' W) to Santa Catarina (27 o 16´ S: 44 o 57´W) states (Fig. 1).
After evaluating the octopus species described by Arocha and Urosa (1982) in the southernmost area of the Caribbean, and papers on the distribution and biogeography of shallow octopus species along the American coast (Voight 1998, Gonçalez et al. unpublished observations), we realized that the south Caribbean Sea is the distribution limit for octopus species with large eggs described from North to Central Atlantic and the Caribbean sea, including P. joubini, P. mercatoris, Octopus briareus Robson, 1929 (egg length 11 15 mm), and Octopus zonatus Voss, 1968 (egg length 6.6 8.2 mm). These species were not recorded in the Amazon reef system or in northeast Brazil (Leite and Haimovici 2006;de Luna Sales et al. 2019), probably because they produce benthic juveniles, with limited dispersal range (Voight 1998, Villaneuva et al. 2016) across long distances and salinity barriers. In addition, the Amazon River mouth might act as a barrier to their dispersal, preventing passage southwards.
The phylogeny indicated that three small species assigned to the Octopus grouped in the Paroctopus clade, suggesting they belong to this genus. The rst species is O. tehuelchus, a small octopus (200 mm ML) with large eggs distributed from southern Brazil to northern Patagonia in Argentina (Jereb et al. 2014). The second is Octopus alecto Berry, 1953, a Paci c pygmy species found in the Gulf of California from Mexico to Ecuador. The third species assigned to the Paroctopus genus is Octopus tchi, another Paci c pygmy species found in shallow waters (down to 30 m) in sandy and muddy substrates from the Gulf of California and Mexico to Ecuador (Jereb et al. 2014).
A recent study using molecular analysis of partial COI gene sequences and traditional morphometry data suggested that O. alecto should be considered Paroctopus (Díaz Santana Iturrios et al. 2019). Additionally, Magallón-Gayón et al. (2019) sequenced the complete mitochondrial genome of O. tchi and pointed out that this species does not belong to the genus Octopus, suggesting that it seems closer to the Paroctopus group. The three species have small size, median to large size eggs and inhabit shallow waters. Our analysis con rms and expands previous studies, suggesting that these species should be reviewed and moved to the Paroctopus genus.

Morphological comparison with related taxa
As noted above in the introduction, Octopus joubini is the name used most frequently for the small egged pygmy species from the North Atlantic Ocean, Caribbean Sea and Mexican Gulf (Jereb et al. 2014). The holotype of this species (BMNH 1889.4.24.30) is 16 mm ML, a female bearing medium ripe eggs measuring 3.2 mm in length (see Table 5). Voss and Toll (1998) further described the species based not only on the holotype, but also on specimens examined by Forsythe and Toll (1991). These last authors observed mature females with 150 3000 ripe eggs of 2.3 4.8 mm in length. Compared to the description of Forsythe and Toll (1991) for O. joubini, P. cthulu sp. nov. has larger eggs (4.2 9.0 mm), a deeper web (20 72 vs 28) and more arm suckers (ASC 102 178 vs 79) (see Table 5). The sole criterion that Norman et al. (2016) used to consider O. joubini a member of the Octopus genus was the "small" size of its eggs. However, considering our genetic and morphological results, we suggest that this criterion needs to be reevaluated.
Compared with the small egged species morphotypes collected from Belize in the Caribbean and deposited on the Santa Barbara Museum (see Table 5), P. cthulu sp. nov. also showed a larger normal sucker index (SDnI 8 13 vs 6 11) and enlarged sucker index (SDeI 10 23 vs 12.5). Another important morphological feature of P. cthulu sp. nov. is the presence of enlarged suckers in two out of three females evaluated, while the morphotype of P. joubini only had enlarged suckers in male specimens P. cthulu sp. nov. had more suckers on normal arms (ASC 102 174 vs 58-94) and also on hectocolized arm (ASCH 56-93 vs 45-70) compared with Belize forms. It also had more gill lamellae (5 6 vs 4), and bigger eggs (4.7 9 vs 3.2).
Another name used for the pygmy octopus from North Atlantic and Gulf of Mexico is O. mercatoris. Pickford (1945) compared O. mercatoris and O. joubini by morphometric indices, and considered the former species as a synonymy of O. joubini. However, Forsythe and Toll (1991) after rearing the two forms of O. joubini (large and small egged) concluded that they are in fact two different species. Their conclusion was based on the hatchling size, as while the small egged specimen produced planktonic paralarvae, the large egged individuals produced benthic juveniles. For these authors, the small egged pygmy species is conspeci c with the holotype of O. joubini, and not the widely studied and better known large egged species, although both species occur in the Caribbean Sea and the Gulf of Mexico. For these authors, the taxonomy of the large egged pygmy species from the northwestern Atlantic is still not clari ed.
Besides the available holotypes and syntypes, we also compared the new species with large egged specimens deposited at National Museum of Natural History (NMNH Smithsonian) from different localities (see Table 5). Our specimens had a larger calamus index (CLI 20 42 vs 21 31.6), shorter spermatophore index (SpLI 40 43 vs 55.5), more suckers on normal arms (ASC 102 174 vs 69 99) and also on hectocolized arm (ASCH 56-93 vs 45) when compared with the large egged morphotypes from South Florida (see Pickford, 1945), and from those in the experiments conducted at the National Resource Center for Cephalopods in Texas(see Forsythe and Hanlon 1980) (Table 5).
In addition, the body pattern when compared to both P. joubini morphotypes is quite different. Paroctopus cthulu sp. nov. species has a characteristic reddish orange coloration, but with variable body patterns that includes also use of the yellow, white and black chromatophores and papillae all across the body. In contrast, O. joubini (small egged) have a dark brownish tone, also described in the large egged morphotype (Forsythe and Toll, 1991) with no ability to modify skin texture other than 3 4 papillae. Mather (1984) also indicated that the O. joubini large egged morphotype became strongly nocturnal after the third week of life, which is compatible with its drab skin and few body patterns, most of them reddish or dark colors.
As the new species is distinct from O. joubini sensu stricto and the large egged morphotype, it must also be compared with other Octopodidae from the southwestern Atlantic, described by Palacio (1977), and more recently by Leite and Haimovici (2006), Vaske Jr and Costa (2011), Haimovici et al. (2009). P. cthulu sp. nov. has the smallest adult size, when compared to all described southwestern Atlantic octopod species (32 mm ML vs 70 mm ML to Amphioctopus burryi Voss, 1950;up to 250 mm ML (Voss, 1951), Octopus americanus Montfort 1802 (Avendaño et al. 2020), con rming that it is the smallest octopod species from southwest Atlantic.
Octopus hummelincki Adam, 1936 has a larger adult size (70 mm ML) and has ocelli on the web under the eyes, has dissimilar spermatophores, ligula, radula and skin color and textures (Burgess 1966;Leite and Haimovici 2006). Amphioctopus burryi is another small octopus that uses gastropods shells and debris as shelters (Hanlon and Hixon 1980). This species as having a complex body pattern, with a grainy skin and a conspicuous purplish brown stripe along the entire leading edge of the arms pairs I to III, which makes its recognition easy. Octopus tehuelchus d'Orbigny, 1835 has a larger adult size (90 mm ML), longer arms with fewer suckers (about 100), and females bear larger eggs up to 18 mm in diameter (Palacio 1977, Voss andToll1998). Pinnoctopus (Callistoctopus) macropus (Risso, 1826) has a distinct larger adult size (150 mm ML, with a distinct red and white coloration on body and arms (Mangold 1998, Leite andHaimovici 2006). Macrotritopus cf. de lippi Verany, 1851 has larger adult size, and, longer and thinner arms, with a skin with pallid color (Mangold 1998), while O. americanus (Avendaño et al, 2020) and O. insularis (Leite et al. 2008) are bigger animals with larger adult size.

Early life stages
The mode of development of octopus hatchlings -whether planktonic or benthic -can often be inferred by morphological traits, involving the body proportions of hatchlings and adults (Boletzky 1974;Boletzky et al. 2002). In general, species producing eggs smaller than 10% of the adult ML, which result in an egg index <10, and smaller hatchlings (hatching size index >5) with short arms (<50 %ML) produce planktonic offspring, while species with large eggs (>10 mm, egg index >10) and large hatchlings with long arms, produce benthic hatchlings. Intermediate sized eggs (6 9 mm), can produce either planktonic or benthic hatchlings (Boletzky 1974, Boletzky et al. 2002, Hochberg et al. 1992).
In P. cthulu sp. nov., eggs ranged from 4.2 to 5.5 mm in length, but larger oocytes (9 mm) were found in mature females, producing an egg index from 14.7 to 28. Hatchlings have a mean ML of 2.52, with a hatching size index ranging from 8.5 to 12.7 and hatchling AL index of 80%. While the egg length suggests either planktonic or benthic hatchlings, the AL index suggests planktonic hatchlings, but the egg index and the hatchling size index strongly indicate benthic hatchlings. Thus, P. cthulu sp.nov. has morphological features and proportions that would t both the planktonic and benthic mode of development.
The peculiar morphological features of P. cthulu sp.nov. hatchlings raise many questions on the nature of its habitat and behavior. Among these features are the large prominent eyes and the robust funnel. The body is fragile and transparent, particularly the arms, with a clear web and a skin lm covering their entire length, with large cavities formed both dorsally and ventrally by the skin lm. As well there is a dense distribution of Kolliker organs on the mantle. These morphological features are typical of planktonic hatchlings instead of benthic ones (Villanueva and Norman, 2018).
Octopus paralarvae and pelagic octopods have both a dorsal and a ventral mantle cavity. In the later, these cavities are believed to facilitate maneuverability, while squid paralarvae have only a ventral mantle cavity (Villanueva and Norman, 2018). The two mantle cavities in Octopus paralarvae might help to increase the hydrostatic pressure inside the mantle cavity, which in turn increases the propulsive jetting and thus displacement of paralarvae, perhaps to balance the lack of ns, which acts as propulsors in squid paralarvae (Vidal et al. 2018). P. cthulu sp.nov. hatchlings have a large continuum dorsal cavity and also both mantle and head ventral cavities formed by the skin lm, suggesting that these cavities might help to increase propulsive jetting and thus swimming performance, while also minimizing sinking. Another strong evidence for this reasoning is the large size of the funnel in relation to the ML of the hatchlings.
It was suggested that E. megalocyathys hatchlings could live in the suprabenthos for a short period of time, which would favor their dispersal before settling de nitively to the benthos (Ortiz et al. 2006). This suprabenthos includes bottom dependent animals, such as mysids, isopods and amphipods, living in the water layer just above the sea oor and performing vertical migrations above the bottom and with good swimming ability (Brunel et al. 1978). Another study on activity, locomotion and behavior of P. joubini has reported that during the rst week after hatching, the young animals are active during the day and in their rst month of life displayed a "semi benthic" behavior, involving moving to higher spots (rocks or edges of the aquaria) and swimming in the open water, often drifting with spread arms in the water column (Mather, 1984). The author also reported that the activity and swimming behavior of P. joubini changed slowly over a few weeks, and adults were only active at night and never moved far from the bottom.
Such behavior of drifting in the water column with spread arms described for young P. joubini would seems also reasonable for P. cthulu sp. nov. hatchlings. That would explain the need for the protuberant eyes and funnel, the arms web; and their lateral extensions, which look similar to swimming keels in the arms of squid paralarvae, besides the large cavities formed by its conspicuous skin lm. The possibility that P. cthulu sp. nov. hatchlings could be temporarily planktonic or suprabenthic, prior to settling to the benthos, is indeed intriguing, as it would indicate a plastic mode of development for octopods, which would combine the advantages of dispersal and large offspring size, and explain the peculiar morphology of P. cthulu sp. nov. hatchlings. This possibility remains open for future behavioral studies.

Habitat and in situ behaviour
The specimens collected at this study were found inside debris on sand/muddy substrate, usually hidden below foliage and branches of terrestrial origin, but not in phanerogam habitats, as O. joubini in the Caribbean and North Atlantic (Eidemiller 1972, Arousa and Urosa 1982, Mather 1982, Tiffany et al.2006).
Paroctopus or pygmy species were reported using gastropod or bivalve shells as their main refuge (Mather 1982a, 1982b, Voight 1990, Iribarne 1990, with eventual use of arti cial dens as shelter (Voight, 1988). The type, size and availability of these shells in uenced the octopus' abundance and possibly fecundity (Mather 1984, Iribarne 1990, Voight 1992. Empty gastropod shells are an important resource for many animals, including octopuses, in shallow benthic marine communities and this dynamic could shape a benthic community structure (McLean 1983). Natural seashells are becoming increasingly scarce in shallow clear and warm waters due to tourism and collection for craftwork and decoration (Souto Alves et al. 2006, Kowalewski et al. 2014, while marine debris is increasingly available due to pollution by debris in the oceans (Jambeck et al. 2015). It is quite possible that P. cthulu sp. nov. nd in this debris an alternative shelter along the beaches of Ilha Grande frequented by tourists.
Considering the consequences of a successful habitat choice for benthic octopuses and the various negative impacts of solid waste on marine ecosystems, it is interesting to see debris as conveying an advantage (see also Anderson et al.1999, Katsanevakis andVerriopolis 2004). This choice of trash has also been observed for other invertebrate species such as hermit crabs (Zuleta 2019) and sea urchins (Barros et al. 2020). This may demonstrate the plasticity that cephalopods show (Hochner et al. 2006;Albertin et al. 2015) and shows the octopuses are adapting to human impact. More studies are being carryied out by our research group to clarify this ecology, which can be important for the conservation of the new species.  Collection  date  1966  1971  1972  2015  2015  2015 fev 2007  2015  2015  2015 fev  2014  Sex  male  male  male  male  male  male  female  female  female  male  female  Maturity  adults  adults  adults  adults  adults  adults  adults  adults  adults  subadults subadu  Total