Out on a limb: novel morphology and position on appendages of two new genera and three new species of ectoparasitic isopods (Epicaridea: Dajidae) infesting isopod and decapod hosts

ABSTRACT Adult female dajid isopods are ectoparasites found on carapaces, within marsupia, or rarely attached to the eyestalks or pereopods of crustacean hosts. We describe two new genera and species whose females exhibit a novel spheroid shape and envelope antennules of hosts: Akrophryxus milvus n. gen., n. sp. parasitizing Ethusa machaera Castro, 2005 (Madagascar, c. 900 m) and Telephryxus clypeus n. gen., n. sp. parasitizing Munidopsis crassa Smith, 1885 (Caribbean, c. 5000 m). Females of both species differ from other dajid species in their spheroid shape and a plate partially surrounding the host antennule; males are distinguished by reduction of pereopods. Females of A. milvus n. gen., n. sp. are distinguished from those of T. clypeus n. gen., n. sp. in possessing antennules and a triangular attachment plate with three holes (T. clypeus n. gen., n. sp. lacks antennules and the broad attachment plate has two medial holes). Males of A. milvus n. gen., n. sp. are distinguished from those of T. clypeus n. gen., n. sp. in having only six pairs of pereopods, rudimentary antennules, single segmented antennae, and pleon fused to pereomere 7 (T. clypeus n. gen., n. sp. with rudimentary seventh pair of pereopods, large single segmented antennules, two segmented antennae, and pleon distinct from pereomere 7). Pre-molt epicaridium larvae and cryptoniscus larvae of T. clypeus n. gen., n. sp. are described. We also describe Aegophila cappa n. sp. found on pereopods of the isopod Aegiochus symmetrica (Richardson, 1905) (Bering Sea, c. 650 m), distinguished from the type species of Aegophila Bresciani, 1966 by oostegite morphology of the female, form of the antennae, and shape of the pleon of males. Tables with comparative characters for genera of Dajidae G. O. Sars, 1883 are provided. Dajid life cycles and development are reviewed.


INTRODUCTION
Attachment to the extremities, including appendages such as antennae and legs, would seem a risky strategy for ectoparasites of aquatic hosts. Although some protozoans and other small ectoparasites and epibionts are commonly found on such host structures or their setae (e.g. Lom 2005; Heiner & Kristensen 2005), macroparasites are rarely attached in such positions. Rather, the bodies of most ectoparasites are closely applied to the bodies of their hosts or within protected regions (e.g. branchial chambers) and often exhibit adaptations (e.g. dorsoventral flattening, modified attachment structures) to withstand the hydrodynamic pressures that would otherwise dislodge them from hosts (e.g. Boxshall 2005a, b;Smit et al. 2014;Baillie et al. 2019). Other unusual sites of attachment are on the eye stalks of their crustacean hosts (Shields & Gómez-Gutiérrez 1996). Crustacean parasites that themselves parasitize other crustaceans are likewise usually found attached to the bodies of their hosts, either on the dorsal or ventral surfaces of the cephalon, the pereon or the pleon; however, there are a few highly modified spheroid or ovate macroparasites, such as certain rhizocephalans, tantulocarids, and copepods, that are known from extremities (e.g. antennules, antennae and pereopods) of their crustacean hosts (Boxshall & Hayes 2019).
Rhizocephalans (Rhizocephala Müller, 1862) are mesoparasitic barnacles that include colonial forms, of which some species (e.g. Thompsonia japonica Häfele, 1911;Thylacoplethus porcellanus Hiller, Williams & Boyko, 2015) can occur on the pereopods as well as the body of the host. An additional species, Thompsonia chuni Häfele, 1911, is known from the eyestalks of its hermit crab host (Häfele 1911;Hiller et al. 2015). Each mature externa (the reproductive portion of the body) of a colonial rhizocephalan is approximately 1-2 mm in length and is attached by a short stalk to an appendage of its host; however, it does not surround the appendage (Hiller et al. 2015). The spheroid externae of these species, like all rhizocephalans, erupt from the body of the host and are connected to an internal root system; they are not originally attached to the appendages (aside from the larval phase that is responsible for injection of the primordial cells and may settle on an appendage prior to the injection; see life cycles in Høeg & Lützen 1995;Baeza et al. 2018).
Tantulocarids (Tantulocarida Boxshall & Lincoln, 1983) are ectoparasites on a range of small crustacean hosts (copepods, peracarids and ostracods) and their life cycle includes sac-like adult female stages attached to the host (both parthenogenetic females and females in the sexual stage are sac-like; see background and life cycles in Huys et al. 1993;Boxshall 2005c;Baeza et al. 2018). Several species of tantulocarid parasites (e.g. Amphitantulus harpiniacheres Boxshall & Vader, 1993;Microdajus tchesunovi Kolbasov & Savchenko, 2009;M. langi Greve, 1965) attach to the antennules, antennae and/or pereopods of their tanaid or amphipod hosts. However, these species have also been found attached to the main body of their hosts (Grygier & Sieg 1988;Boxshall & Vader 1993;Kolbasov & Savchenko 2009). There are a few tantulocarid species (e.g. Microdajus aporosus Grygier & Sieg, 1988) that have only been reported from appendages (pereopods) of tanaids (Grygier & Sieg 1988). Species of Microdajus Greve, 1965 are minute (< 1 mm in length) and are attached by a stalk to the appendage of their hosts (Boxshall & Hayes 2019;Boyko et al. in press). ZOOSYSTEMA • 2021 • 43 (4) The only other previously known crustacean parasites that may be appendage specialists are some species of mesoparasitic nicothoid copepods (Nicothoidae Dana, 1852;see review in Kakui 2016). In particular, two species (Nicorhiza humesi Lincoln &Boxshall, 1983 andN. rockallensis Lincoln &Boxshall, 1983) are found attached to the bases of pereopods 3-6 of their asellote isopod hosts (Lincoln & Boxshall 1983). Adult females of Nicorhiza Lincoln & Boxshall, 1983 and other genera in the "Rhizorhina group" (sensu Boxshall & Halsey 2004) superficially resemble rhizocephalans, including convergently possessing an absorptive rootlet system within the host. These copepods are spheroid, relatively small (400-600 µm in length) and attach to hosts by a modified oral cone. The point of attachment to hosts for some nicothoids (e.g. species of Diexanthema Ritchie, 1975;see Boxshall & Harrison 1988) remains unknown and other species might be found to vary in their position on hosts (sampling has been low for many) but none are known to surround the appendages of their hosts.
Among the parasitic isopods (Epicaridea Latreille, 1825), which likewise exclusively parasitize other crustaceans, there were no species previously found attached to the antennules and only one known to attach to pereopods of their hosts. Epicaridea contains two superfamilies (Bopyroidea Rafinesque, 1815 andCryptoniscoidea Kossmann, 1880), species of which nearly all externally parasitize calanoid copepod hosts as their intermediate hosts and then parasitize one or more of a wide range of crustaceans (cirripedes, decapods, euphausiids, mysids, ostracods, peracarids) as definitive hosts. The most diverse family within Bopyroidea is Bopyridae Rafinesque, 1815 with over 600 species (Boyko et al. 2008a onwards), which are mostly ectoparasites in the branchial chambers of their definitive hosts, although some are also found attached to the ventral surfaces of host abdomens and a very few occur on other parts of hosts or are endoparasites (Trilles 1999;Williams & Boyko 2012). Females of most bopyrid species have a dorsoventrally flattened form, as do those of species in Ionidae H. Milne Edwards, 1840 which are exclusively branchial chamber parasites (see Boyko et al. 2013), whereas females of species in Entoniscidae Kossmann, 1881 are endoparasitic and vermiform. Among the cryptoniscoids, there are many species in which sexually mature females exhibit a sac-like form (Trilles 1999), but they are irregular in shape, do not form a spheroid, and are typically found in the brood chamber or shell (for the case of barnacles) of hosts as primary parasites or hyperparasites of a wide range of hosts (Williams & Boyko 2012).
Within Cryptoniscoidea the family Dajidae G. O. Sars, 1883(Boyko et al. 2013 presently contains 56 species that typically possess a cryptoniscus larval stage with an oral sucker (see more in Remarks under Telephryxus clypeus n. gen., n. sp.) used to attach to the definitive host (see Gilson 1909;Coyle & Mueller 1981;Gómez-Gutiérrez et al. 2017). Most species in this group are attached as adults to the dorsal carapace of their hosts, typically on caridean shrimp and euphausiids or occur in the marsupia of mysids (see G. O. Sars 1898;Wasmer 1988;Shimomura et al. 2005;Gómez-Gutiérrez et al. 2017;de Almeida Alves-Júnior et al. 2019). One species (Oculophryxus bicaulis Shields & Gómez-Gutiérrez 1996) uses modified antennae to attach to the eyestalk peduncle of euphausiid hosts, whereas species of Heterophryxus G. O. Sars, 1885 also attach to eyestalks but do so with their fifth pereopods. However, species of Heterophryxus and Oculophryxus Shields & Gómez-Gutiérrez 1996 retain the dorsoventrally flattened typical dajid form as adult females and are closely applied to the host body. Aegophila socialis Bresciani, 1966 is the sole dajid as well as epicaridean known to parasitize the pereopods of its host, a species of aegid isopod; the body of females is more flattened than is typical of other dajids but is otherwise quite similar. Notophryxus lateralis G. O. Sars, 1885, is found attached to an appendage, either the endopod of the sixth pereopod or the sixth gill (G. O. Sars 1885;Field 1969;Gómez-Gutiérrez et al. 2017) but this species is in a very protected position on the ventral surface of its host.
Herein, we report on two new species of dajids that exhibit an unusual spheroid morphology and novel attachment position partially surrounding the antennules of their decapod hosts. In contrast to the highly modified female morphology, the males, epicaridium and cryptoniscus larvae of these species are typical of dajids in gross morphological details. One species was collected from the western Indian Ocean off Madagascar in nearly 1000 m depth, whereas the other was obtained in exceptionally deep (c. 5000 m) waters from the southern Caribbean. We also describe a third species of dajid, a new species of the genus Aegophila Bresciani, 1966 from the Bering Sea, found attached to a pereopod of an aegid isopod host. Although the morphology of females of this species is atypical of dajids and apparently specialized for attachment to the legs of hosts, females of this species do not exhibit the spheroid shape found in females of the two new genera attached to host antennules. The present study thus expands our knowledge of the morphological diversity and host attachment strategies within Dajidae, as well as parasitic isopods in general.

MATERIAL AND METHODS
Methods of specimen collection are indicated in the "Type Material" sections below; data on collections (e.g. longitude and latitude coordinates, depth) are provided from the original records where possible. Carapace length (CL) of hosts was measured from the anterior end of the rostrum to the posterior edge of the carapace and carapace width (CW) was measured as the maximum width of carapace, using calipers. Isopod parasite size is given as maximal diameter for females of the two new genera (Akrophryxus n. gen. and Telephryxus n. gen.) and maximal length for all males and females of Aegophila (from anterior border of cephalon to terminus of uropods). All measurements were made with a micro-scale tool, ocular micrometer, or from drawing tube sketches.
Line drawings of parasites were made by sketching specimens using drawing tubes attached to Olympus compound and dissecting microscopes and using Adobe Illustrator to trace original sketches and produce final figures. Light micrographs were created with a Macropod Pro kit (MacroscopicSolutions) and resulting pictures were aligned and stacked with the focus stacking software Zerene Stacker (10-65 images from bottom to top of specimens). For Scanning Electron Microscopy (SEM) preparation of epicaridium larvae, specimens were dehydrated in an ascending ethanol (EtOH) series, starting with 70% EtOH and ending with 100% EtOH. Specimens were then dried in a Samdri 795 Critical Point Dryer, mounted on aluminum stubs, coated with gold using an EMS-550 Sputter coater, and viewed with a FEI Quanta 250 SEM.
All specimens were deposited in the Museum National d'Histoire Naturelle, Paris, (MNHN) and National Museum of Natural History, Smithsonian Institution, Washington D.C., (USNM). References are provided for taxonomic authorities of parasite taxa but not for those of hosts. eTymology. -The species name is derived from the Latin for cloak or cape (cappa), in reference to the fused oostegite 5 and lateral body wall of the female's resemblance to a cloak drawn around the body. The gender is feminine.

SYSTEMATICS
DisTribuTion. -Known only from the type locality and type host.

remarks
The new species is clearly conspecific with Aegophila socialis; females of both species show dorsoventral compression of the body, lack of antennules and antennae, expanded, clublike mandibles, the same arrangement of oostegites 1-4, the presence of a rounded knob and fringe of thin papillae on the mediovental margin of oostegite 5, as well as fusion of the fifth oostegites with the ventral body wall, and the pleon being a small, unsegmented protuberance. Males of the two species are similar in having fusion of the cephalon with the first pereomere and presence of rounded posterolateral lobes corresponding to the side of pereomere 1, the antennules as unsegmented flaps fringed by setae, and the pleon with all segments fused and lacking pleopods. Diagnostic differences between females of the two species include: oostegite 1 large, ovate in A. cappa n. sp. (small, "spoon-like" in A. socialis), oostegite 2 narrow in A. cappa n. sp. (broad in A. socialis), and oostegite 4 subquadrate in A. cappa n. sp. (ovate in A. socialis). Bresciani (1966) described A. socialis as lacking maxillipeds, but they are clearly present in A. cappa n. sp. and it is unclear if he overlooked them or if they are truly lacking in A. socialis; the latter appears unlikely since maxillipeds are required for oxygenation of the brood (Gilson 1909;Cericola & Williams 2015). Bresciani (1966) described the pereopods of A. socialis as "small and badly segmented, and do not show any special characters" but did not illustrate them well. In A. cappa n. sp., the pereopods have highly recurved dactyli and show fusion only of the propodus and carpus, with the carpus inner margin being expanded as a cup for insertion of the dactylus.
The males of the two species differ in the form of the antennae (multisegmented with a single flagellum in A. cappa n. sp. vs a single segment with a single flagellum in A. socialis) and the shape of the pleon (much longer than wide in A. cappa n. sp. vs length and width subequal in A. socialis). Bresciani (1966) did not describe the pereopods in detail but did illustrate them. His illustrations appear to show a difference in pereopods 1 and 2 (rounded propodus with typically recurved dactylus) vs 3-7 (elongate propodus and highly recurved dactylus) that is also seen in A. cappa n. sp. However, he drew pereopod 1 as having the propodus and carpus fused whereas in A. cappa n. sp., they are distinct.
It is not clear which specimen of Aegiochus symmetrica was the host of A. cappa n. sp. as all the of the potential host aegids collected at the parasite's type locality (Sta. 4772, USNM 39293) have their legs intact and no evidence of damage is visible; however, a label reading "1 sent to W. W. Alpator / Jan 1923" is present in the jar and it is possible that this gifted specimen was the (now lost) host specimen of the holotype of A. cappa n. sp. eTymology. -The genus name is derived from the Greek prefix άκρο-(ákro-), meaning "limb" or "extremity" in reference to the specialization of the female of the species on the antennule of the host crab rather than on the host body, in combination with -phryxus, from the Boeotian prince Phryxus (Φρίξος) of Greek mythology, a common ending for dajid, as well as some bopyrid, genera. The gender is masculine.
remarks See Remarks below under Akrophryxus milvus n. gen., n. sp. for comparison of female and male characters of Akrophryxus n. gen., Telephryxus n. gen., and other dajid genera. eTymology. -The species name is derived from the Latin for "kite" in reference to the plate partially surrounding the host antennule that is reminiscent of the shape of a kite shield, defensive armor used primarily in Western Europe in the 10th-13th centuries. The gender is masculine.
DisTribuTion. -Known only from the type locality and type host.

remarks
The female of Akrophryxus milvus n. gen., n. sp. is very different from those in all other dajids, excepting Telephryxus clypeus n. gen., n. sp. (Table 1). The spheroid shape of the body and possession of a plate partially surrounding the host antennule are shared only by these two genera. The male of A. milvus n. gen., n. sp. is likewise very different from those in all other dajid genera (Table 2), including T. clypeus n. gen., n. sp. in having only six pairs of pereopods with no trace of the seventh pair (pereopod 7 of two elongate segments in the sole species of Oculophryxus and of one or two stub-like segments in T. clypeus n. gen., n. sp.). The male of A. milvus n. gen., n. sp. also differs from those in most genera in having the pleon fused with pereomere 7 (distinct from pereomere 7 in males in all other genera except Holophryxus Richardson, 1905 and Oculophryxus), and from all other genera in having a reduced antennule and an antenna with a single large segment (males of other genera with antennules of at least one well-developed segment and all with antennae segmented and with distal flagella). The female of A. milvus n. gen., n. sp. can be distinguished from that of T. clypeus n. gen., n. sp. in lacking antennules and in having three medial holes on the narrow, triangular attachment plate partially surrounding the host antennule with no additional lateral holes (T. clypeus n. gen., n. sp. has flattened antennules and a broad, subquadrate attachment plate partially surrounding the host antennule and having two large medial holes, the largest with two additional small lateral holes). The male of A. milvus n. gen., n. sp. can be distinguished from that of T. clypeus n. gen., n. sp. in having only six pairs of pereopods (six fully-developed pairs plus a rudimentary seventh pair of one or two small segments in T. clypeus n. gen., n. sp.), a reduced antennule and a single segmented antenna (large ovate single segmented antennule and two segmented short antennae in T. clypeus n. gen., n. sp.), and the short and rounded pleon fused to pereomere 7 (pleon distinct from pereomere 7, elongate and tapered pleon in T. clypeus n. gen., n. sp.).
eTymology. -The genus name is derived from the Greek prefix τῆλε-(tele-), meaning "far away" in reference to the specialization of the female of the species on the antennule of the host crab rather than on the host body (as in "far away from the host's body") as well as to the occurrence of the sole species in the abyssopelagic zone (far away from the habitat of most dajids), in combination with -phryxus, from the Boeotian prince Phryxus (Φρίξος) of Greek mythology, a common ending for dajid, as well as some bopyrid, genera. The gender is masculine. Diagnosis. -Female. Body spheroid; cephalon externally indistinguishable from pereon. Antennules each as flat triangular plate; antennae each as flat rectangular plate lateral to oral cone. Maxillipeds inflated, rectangular, each with recurved maxilliped digitiform extension ("appendix") at posterolateral corner. Pereopods 1-5 subequal in size and shape. Oostegite 1 largest, broadly ovate with large triangular posterior accessory lobe; oostegites 2-4 flat, ovate; oostegite 5 flat, tapering. Pleon apparently modified as subquadrate, broad, thickened plate with two circular medial holes, subequal in size; hole surrounding distal portion of host antennular peduncle with one small additional hole on each lateral margin. Male. Cephalon fused with pereomere 1; cephalic slits present. Antennules each as single ovate plate lateral to and extending posterior to oral cone; antennae of two segments each, flagella absent. Pereomeres 2-7 distinct, lateral margins recurved ventrally. Pereopods 1-6 subequal in size and shape; pereopod 7 reduced, single rounded stub on right side, rounded stub plus dactylus on left side. Pleon tapering to rounded tip, all segments fused, distinct from pereomere 7; pleopods lacking. Cryptoniscus larva: Body tear-drop shaped. Cephalon medial region of posterior margin convex, lateral regions concave, posterolateral margins extended posteriorly. Antennules of three articles each, basal article triangular, article 2 quadrate, article 3 digitiform, inserted into article 2 distoventrally. Antennae of nine articles each, peduncular articles subequal in size; flagellar articles approximately half width of peduncular articles. Oral cone lacking oral sucker (possible artifact due to loss during collection). Pereomeres 1-7 with entire coxal plates. Pereopod 1 with short, slightly curved dactylus, propodus semi-spherical; pereopods 2-6 with more elongate curved dactyli, propodi progressively more elongate. Pereopod 7 with long, slightly curved dactylus, propodus elongate with two large multifid setae near base of dactylus; carpus with distal large multifid seta; merus triangular with anterodorsal edge extending into spine-like extension, ischium triangular, basis cylindrical. Pleon with five pairs of biramous pleopods, endopods cylindrical, exopods triangular. Pleotelson oval. Uropods biramous, sympod broad, endopod slightly longer than exopod. remarks See Remarks above under Akrophryxus milvus n. gen., n. sp. for comparison of female and male characters of Akrophryxus n. gen., Telephryxus n. gen., and other dajid genera. eTymology. -The species name is derived from the Latin for "shield" in reference to the broad subquadrate plate that partially surrounds the host antennule and is reminiscent of the shape of a shield. The gender is masculine.
DisTribuTion. -Known only from the type locality and type host.
remarks See Remarks under Akrophryxus milvus n. gen., n. sp. for comparison with Telephryxus clypeus n. gen., n. sp. and other dajids. The maxilliped digitiform extension ("appendix" of Rustad 1935) appears homologous with the lateral lobes of the barbula found in bopyrids (Markham 1985). As in bopyrids, the maxillipeds are used to pump water through the brood chamber, oxygenating the eggs or larvae (Gilson 1909;Cericola & Williams 2015). However, in the case of females of species in the two new genera, the digitiform extension of the maxilliped also fits into the groove of oostegite 1, thus further aiding in oxygenation by moving the whole first oostegite. It may also aid in maintaining larvae in the brood chamber prior to release (Cericola & Williams 2015). The epicaridium larvae of Telephryxus clypeus n. gen., n. sp. were in pre-molt, having a wrinkled cuticle (particularly evident surrounding the cephalon, antennae and pereopods) similar to that shown in larvae of other dajid species (G. O. Sars 1898: pl. 94;Gilson 1909: figs 10, 11;Shimomura et al. 2005: fig. 12). As discussed by Gilson (1909), there are two stages of epicaridium larvae: the first is maintained within the brood chamber of the female and the second is released from the female as the planktonic phase prior to attachment to a copepod intermediate host (Fig. 11A-G). In addition to the yolk provisioned endogenously to the epicaridium larvae, they may also be provisioned with exogenous sources of nutrition taken up across the embryonic epithelial layer (as discussed by Strömberg [1971] and shown to occur in cryptoniscoids by Goudeau [1977]) that fuel their final development prior to release into the water column. After molting, the second stage (mature) epicaridium larvae of dajids show distinct segmentation of appendages (e.g. G. O. Sars 1898;Gilson 1909;Tattersal 1911;Taberly 1957a). The pleopods of the epicaridium larvae of T. clypeus n. gen., n. sp. all appear biramous (as they do in other pre-molt larvae previously described; e.g. Stephensen [1913], Shimomura et al. [2005]). However, it is not clear if any or all of these will remain biramous or become uniramous in the mature epicaridium larvae. Pleopods of mature epicaridium larvae of dajids have been described with both uniramous and biramous states: pleopods 1-5 uniramous (Gilson 1909) or pleopods 1-4 biramous and 5 uniramous (Taberly 1957a;Coyle & Mueller 1981).
The uropods of the pre-molt epicaridium larvae of T. clypeus n. gen., n. sp. are biramous, as are those of the mature epicaridium larvae described by Taberly (1957a). However, the orientation of the endopod and exopod of the uropods is ninety degrees rotated ("superposées" of Trilles 1999) from  Sars 1898; Gilson 1909) but, as suggested by Taberly (1957a), these larvae may not have been observed in the proper orientation. Tattersall (1911) considered uniramous pleopods and uniramous uropods a defining characteristic for dajids; however, the key of Bourdon in Trilles (1999) is probably more accurate for dajid characters (pleopods and uropods both biramous). The cryptoniscus larva of T. clypeus n. gen., n. sp. lacks an oral sucker, a characteristic found in most dajid species for which larvae have been examined (Fig. 11J). However, we think this represents a case where the attachment structure fell off prior to or during collection. As noted by Taberly (1957b), the oral sucker of dajids is easily detached. Other records of dajid cryptoniscus larvae lacking an oral sucker (e.g. Holophryxus alaskensis Richardson, 1905, Zonophryxus quinquidens Barnard, 1914 should be considered as possible instances where this structure was detached. Although Coyle & Mueller (1981) collected several cryptoniscus larvae of H. alaskensis and reported they lacked an oral sucker, we suggest collection of new material of all such species is needed to confirm absence or presence of the oral sucker in cryptoniscus larvae across dajid taxa.  (Table 2); although they are found inside the brood chambers of their female partners, as are the males of many other dajid species, they are essentially entombed by the fusion of the female's pereomeres and pleon around the antennae of hosts. Although the females of both new genera are superficially similar in many aspects (Table 1), the males show distinct, generic-level differences ( Table 2). The similarity of the females' spheroid morphology could either be due to a shared common ancestor or to convergent adaptations for parasitizing antennules of their hosts. More specimens and, ideally, molecular data would be needed to address these possibilities. Dajid life cyles, such as that of Prodajus ostendensis Gilson, 1909 from the opossum shrimp host Gastrosaccus spinifer (Goës, 1864), show that the female parasites reach sexual maturity and release epicaridium larvae that seek out a calanoid copepod, pass through a microniscus larval stage that feeds on copepod hemolymph, and ultimately transform into a cryptoniscus larva that seeks out and settles on a definitive host (Fig. 11A-G). The cryptoniscus larva likely settles on the host (possibly in the branchial chamber) and then moves to its final destination (see Coyle & Mueller 1981). Unlike in other epicaridean parasites that exhibit protandric hermaphroditism (Williams & Boyko 2012), the cryptoniscus larvae of dajids are supposedly of separate sexes and males do not turn into females (Gilson 1909); larvae either become juvenile females (Fig. 11F) or males (Fig. 11G); this, however, needs more study for confirmation. Males move onto females, eventually entering the brood chamber and fertilizing the eggs of sexually mature females (Fig. 11A, H, I). Some researchers (Gómez-Gutiérrez et al. 2010;2017) have suggested the cryptoniscus larvae of dajids can act as micropredators (feeding transiently on multiple hosts) but this has not been definitively demonstrated.  et al. (2007) stated "both of these parasites [dajid and copepod] utilise the lumen of the mysid marsupium as a microhabitat and its eggs (and embryos) as food (Ohtsuka et al. 2005;Shimomura et al. 2005)" but, once again, there is no data given in Ohtsuka et al. (2007) or either of the papers cited therein to show that dajids actually consume eggs. The sentence in Ohtsuka et al. (2007) "No host eggs remained when a fully mature female isopod was present within the marsupium, suggesting that all eggs in a brood were consumed by the isopod" indicates that these authors were speculating on the reason for the lack of host eggs and did not actually observe any predation by the dajids and that the copepod is the only egg predator in this relationship. We are unaware of any published data showing that dajids consume host eggs, although some cryptoniscoids are known to do so (reviewed in Buhl-Mortensen et al. 2020).
We propose the hypothetical life cycle and developmental transformation as the pathway by which females of Akrophryxus milvus n. gen., n. sp. and Telephryxus clypeus n. gen., n. sp. come to exhibit their spheroid morphology. In both new taxa, the life cycle is similar to that described above for dajids in general ( Fig. 11A-G). However, the cryptoniscus larvae (Fig. 11N) either settle in the branchial chamber and then move onto the antennule or they may settle directly on an antennule. After attachment using their pereopods (and possibly the oral sucker) to the antennule, the body of a cryptoniscus larva that is transitioning to a juvenile female begins to dorsoventrally flex and the sides of the body expand (Fig. 11O). Eventually the pleon meets the anterior end and the two coalesce, as do the sides of the body (Fig. 11P). Thus, the plate (or "shield") partially surrounding the host antennule appears to be a highly modified pleon which gradually surrounds the host antennule during development of the parasite (Fig. 11P-R). Alternatively, the plate could be an anterior extension of the cephalon or sides of the body. These appear less likely scenarios considering the shape and orientation of the features. The holes in the plate (Figs 4; 7) represent areas where complete fusion between the cephalon, pleon, and sides of the body has not occurred yet and it is through these openings that the epicaridium larvae of the dajids are likely released. The males reside in the brood chamber of females (as seen in many other dajids although, as mentioned, they are trapped by the fusion of the female body) and fertilize the brood (Gilson 1909). The feeding biology and nutrition of males remains unknown; they may be non-feeding and derive nutrition from stores remaining from the cryptoniscus larval stage.
The male and larval characters (epicaridium and cryptoniscus larvae) display characteristic dajid features. However, the males are unique in their degree of posterior pereopod reduction: lacking a seventh pereopod (Akrophryxus milvus n. gen., n. sp.) or seventh pereopod reduced to one or two stub-like segments (Telephryxus clypeus n. gen., n. sp.), a feature previously recorded in only one other dajid (Oculophryxus bicaulis, with the seventh pereopod reduced to two elongate segments). Finally, the cryptoniscus larval specimen of T. clypeus n. gen., n. sp. did not possess an oral sucker, but we suggest this is likely an artifact due to loss of the structure during handling as the suckers are fragile and break off easily (see Taberly 1957b). Future research should examine the life history of these species to document the changes in form of the female through development to sexual maturity and verify the ontogenetic origin of the plate partially surrounding the host antennule (see Fig. 11N-R), as well as clear up remaining questions on larval characters, such as the presence or absence of a sucker in the cryptoniscus stage and the characters of molted (second stage) epicaridium larvae.
The abyssopelagic specimens of T. clypeus n. gen., n. sp., being found on a host collected from 4956-4997 m, are the deepest known of any dajid species; the prior record was the type specimens of Arthrophryxus beringanus Richardson, 1908 Due to the fact that these are parasites of deep-water hosts, it is unlikely that any questions regarding ontogeny can be tested experimentally (i.e., via rearing). Instead, these questions will most likely need to be addressed with newly fixed material of these dajids in their various larval and developmental stages as obtained from other studies. We therefore urge researchers to retain decapods with unidentified parasites, particularly those found on antennules, properly fix them for morphological and molecular studies and, most critically, deposit them in museums for later analysis.