Revision of the genus Acryptolaria Norman, 1875 (Cnidaria, Hydrozoa, Lafoeidae)

The genus Acryptolaria is reviewed, with a complete redescription of the type specimens, with the exception of A. andersoni and A. rectangularis for which type material could not be located. The genus includes 16 valid species, though A. andersoni is insufficiently known. All records found in the literature have been checked. The cnidome proved to be a useful tool for species identification. A key for the identification of the species of the genus is also presented.


Introduction
Acryptolaria Norman, 1875 is an easily recognizable genus of deep-sea benthic hydrozoans, but identification at the specific level is a difficult task. Many authors (e.g. Allman 1888; Ritchie 1911) have emphasized the paucity of characters for species differentiation in this genus. This, together with the poor original descriptions of many nominal species, makes identification at the species level almost impossible. The present taxonomic knowledge on the genus has become an impediment for any study including species of Acryptolaria. In order to solve this problem, the revision of the type material of the known species of Acryptolaria has been carried out. In the present study, type material of most of the species has been examined and a redescription of the species is provided, together with new figures. Also included is information concerning the cnidome, which proved to be a useful tool for the identification of the species.

Material and methods
The material studied belongs to collections of the Institut Royal des Sciences Naturelles de Belgique, Brussels, Belgium (IRSNB), the Natural History Museum, London, UK The erect colonies of Lafoeidae may be divided into two basic kinds, according to the arrangement of the fascicular tubes and their branching. Although these two types were first noted by Allman (1888), subsequent authors gave little or no attention to their distinction (Marques et al. 2006). Acryptolaria differs from Grammaria in this character, the colonies of the former having a central axis from which all primary branches arise, the fasciculation being complemented by additional tubules with no branching. In Grammaria, however, the fasciculation is composed of many tubes from any of which the primary branches arise.
Although we consider that two categories of nematocysts characterized the species of Acryptolaria (large putative macrobasic mastigophores and small putative microbasic mastigophore), only one category could be observed in a few species because of the almost complete absence of coenosarc in the type material examined and/or all the material was mounted in microslides.
As shown above, many species currently assigned to Acryptolaria were originally described in the genus Cryptolaria, but both genera are easily distinguishable by the presence of nematothecae in the latter.

Material examined
Lectotype: MCZ 9063, Atlantic Ocean, Gulf Stream Expedition, several fragments up to 20 mm long. Paralectotype: MCZ 9077, Atlantic Ocean, Gulf Stream Expedition, four fragments up to 18 mm long and one incipient stem ca 5 mm high. Other material: MCZ 45714, off Double-Headed Shot Key (23u579300 S, 80u299150 W), 315 fathoms (567 m), a vial with four fragments up to 9 mm included in the type material of Acryptolaria longitheca.

Description
''Hydrocaulus attaining a height of about two inches [50.8 mm], irregularly branched, with a pinnate disposition of the ultimate ramuli'' (Allman 1877, p 20). Presently, lectotype material composed of several fragments up to 20 mm in length. Most fragments belonging to the strongly polysiphonic basal part of stem. Stem hydrothecae barely visible. Stem straight, giving rise to lateral branches at right angles, probably following an alternate pattern. Branches straight.
Hydrothecae alternately arranged in approximately one plane. Hydrotheca almost cylindrical; diameter sharply decreasing at basal part and, sometimes, slightly from the middle to the distal part. Hydrotheca gently curved abcaulinarly; adcauline wall approximately two-thirds adnate to internode. Adcauline wall convex; abcauline wall practically straight at basal half, concave at distal half. Hydrothecal aperture circular, parallel to longitudinal axis of branch or slightly directed upwards or downwards. Sometimes adcauline wall longer than abcauline one and, consequently, hydrothecal aperture oblique. Rim even, sometimes with few short renovations.

Remarks
The material described by Allman (1877) is presently divided into two lots, both held in the Museum of Comparative Zoology at Harvard University and, therefore, should be considered as syntypes of the species. We chose the material MCZ 9063 as the lectotype and MCZ 9077 as paralectotype of A. abies (International Commission on Zoological Nomenclature 1999, Article 74). Allman (1877, p 20) already noticed that ''the hydrothecae of this species [A. abies] are considerably smaller than those of either C. conferta or C. longitheca. They can be seen … to be of a very different shape from those of the two former species, being here of an elongated flask-shape, tumid below and gradually narrowing towards its orifice''. Allman (1877, p 20) also pointed out that ''Here [A. abies] also they [the hydrothecae] are plainly provided with a distinct floor, and in all respects resemble a typical sertularian hydrotheca'', and figured (Plate 13 Figure 2) the hydrothecae completely closed at their base. Nevertheless, the examination of type material leaves no doubt that the hydrotheca is devoid of a floor or diaphragm, as shown in Figure 1. The sharp decrease in diameter of the hydrothecae at the hydrothecal base could account for Allman's (1877) mistake.
In the examined type material of A. longitheca (MCZ 45714), there was a small vial with four fragments up to 9 mm long which clearly belong to A. abies. In these fragments the branching is alternate in one plane and the angle between stem and branches is ca 45u. The shape and size of the hydrotheca is in complete agreement with the type material of A. abies.

Distribution
According to Allman (1877), the label belonging to the type material of A. abies was effaced, it not being possible to ascertain the exact locality and depth from which the specimen came. The material of A. longitheca here assigned to A. abies was collected at a depth of 567 m off Double-Headed Shot Key (23u579300 S, 80u299150 W), Florida.

Remarks
As stated above, the type material of A. andersoni, type species of the genus Acryptolaria, is presumably lost. Although there is little doubt that Totton's species actually belongs to Acryptolaria, it is impossible to assign any material to it without examining its type.
As indicated before, Norman (1875) did not recognize the coppinia of A. andersoni, wrongly describing it as a new species of hydroid (Scapus tubulifer).
Figure 2. Acryptolaria angulata (Bale, 1914). (A-C) Hydrothecae. All drawings from the holotype. Scale bar: 400 mm. Hydrothecae alternately arranged approximately in one plane. Hydrotheca mostly cylindrical, but with two distinct constrictions due to sharp perisarc invaginations, narrowing hydrothecal diameter which, in any case, is smaller at the adnate part; diameter also decreasing at basal part. Hydrotheca curved twice: strongly abcaulinarly at distal part of adnate portion and slightly upwards just after the adnate wall becomes free. Abcauline wall straight but strongly inflected at approximately half of its length and provided with a sharp invagination at that point; basal half parallel to longitudinal axis of branch, distal half forming an angle of ca 70u with that axis. Adcauline wall adnate for approximately half its length, though hydrothecae with free part longer or shorter than adnate part also present. Adcauline wall convex at adnate part, with a strong invagination at approximately half of its length. Hydrothecal aperture circular, oblique, upwardly directed. Rim even, sometimes with several short renovations.

Remarks
All material examined consisted of microslide preparations, making it difficult to find nematocysts. However, at least one category of nematocyst is present.
Although the material studied is in three whole-mount preparations, we believe it belongs to a single colony, as there is no indication that the original description (Bale 1914) was based on more than one colony. Therefore, we considered the three slides as parts of the holotype. Bale (1914, p 166) characterized this species by the ''sharply geniculate outward bend, with the noticeable crease in the angle, also by the distinct boss projecting into the hydrothecal cavity, opposite the crease. Immediately after the bend the axis of the hydrotheca is at a right angle with that of the lower portion, but it usually bends a little upward again, so that the distal half is not horizontal but slightly ascending''. The distinct adcauline invagination of perisarc was also found by Vervoort (1966) examining a slidemounted specimen, which he believed to be Bale's schizosyntype or schizoparatype, and by Schuchert (2003), who also examined type material of Acryptolaria angulata (three slides belonging to syntype).
Although the species A. angulata, A. bulbosa, and A. rectangularis share some similarities, we regard them as valid. However, confusion concerning these species may be found in the literature. At least two records of A. angulata refer to different species: Hirohito's (1995) material, lacking the adcauline invagination, is here considered referable to A. bulbosa (see below), and the material described by Vervoort (1966) is here considered as A. rectangularis (see below).

Distribution
Acryptolaria angulata has been collected at depths from 180 (Bale 1914) to 324 m (Bale 1915) on bottoms of sandy mud and small stones (Schuchert 2003). At present, it is known only from the Great Australian Bight (Bale 1914(Bale , 1915 and the Kei Islands, Indonesia (Schuchert 2003).
Hydrothecae alternately arranged, approximately in one plane. Hydrotheca mostly cylindrical; though diameter smaller at adnate portion and decreasing at basal part. Hydrotheca curved twice; strongly outwards at approximately half its length and later slightly upwards, clearly far from the point at which the adcauline wall becomes free. Adcauline wall slightly convex at basal part, extremely convex at approximately half its length and concave at distal third. Abcauline wall straight or slightly convex at basal and distal third, but provided with a sharp invagination at approximately half its length, forming a sort of hump. Approximately half of adcauline wall adnate to internode, though hydrothecae with free part either longer or shorter than adnate one also present. Hydrothecal aperture circular, oblique, upwardly directed. Rim even, sometimes with up to four short renovations.

Additional information
Hirohito (1995, p 102) described the gonosome of this species (as A. angulata), noting that the bottle-shaped gonothecae are aggregated into a coppinia, that the gonophores are in the form of fixed sporosacs, producing acrocysts, and that there are modified hydrothecae diverging and entangled with each other, forming a canopy.

Remarks
Acryptolaria bulbosa has been considered conspecific with A. angulata by several authors (e.g. Hirohito 1995; Vervoort and Watson 2003). Stechow (1932) considered that the species should be called Cryptolaria angulata var. bulbosa if transitions were found. Nevertheless, we consider both species as valid. Although the general appearance of the hydrothecae is similar, only A. angulata is characterized by the presence of the adcauline boss projecting into the hydrothecal cavity. Moreover, we have found clear differences in the size of the larger nematocysts in the cnidome of the type material of A. angulata and A. bulbosa (7.7-8.864.4-5.5 mm and 21-2366-7 mm, respectively). There are also slight differences in the shape of the hydrotheca. In A. angulata the upward bending of the hydrotheca begins almost immediately after the adcauline wall becomes free, whereas in A. bulbosa it starts distinctly further, with a relatively long free portion of the adcauline wall forming an angle more or less perpendicular to the longitudinal axis of the branch. The abcauline side of the abcauline perisarc invagination is distinctly sharper in A. bulbosa, and the abcauline wall forms a much marked hump after it. Furthermore, the size of the hydrotheca in A. angulata seems to be larger, especially concerning the length of the hydrotheca (cf. Table XV). The material from Sagami Bay described as Acryptolaria angulata by Hirohito (1995, p 102), who considered Cryptolaria angulata, C. rectangularis, and C. bulbosa conspecific, clearly does not belong to Bale's species, as it lacks the adcauline invagination of the hydrothecal wall. The general shape of the hydrotheca conforms better with A. bulbosa. The material described by Gravier-Bonnet (1979) as A. rectangularis is different from that described by  and Millard (1967, in part), and it is quite similar to A. bulbosa both in shape and size of the hydrotheca.
The material assigned to A. rectangularis by Millard (1967) apparently includes two species: A. rectangularis (ABD 15P) and A. bulbosa (AFR 1251D). Millard indicated that ''the two samples have hydrothecae of very different sizes, though the proportions are similar''. Moreover, it is possible to recognize both species in Millard's (1967) figures. If these records can be confirmed, the southwestern Indian Ocean would be an area of sympatry of A. rectangularis and A. bulbosa.
The material assigned to Acryptolaria rectangularis by  conforms more to the concept of A. bulbosa adopted here, particularly in the shape and the size of the hydrothecae. In A. rectangularis the free portion of the adcauline wall is straight, whereas in A. bulbosa it is concave and distinctly bends upwards, as described by Millard. Therefore, the two coppiniae described by her would be, actually, the still undescribed coppiniae of A. bulbosa.
The same applies to  material of Acryptolaria rectangularis, which could belong to A. bulbosa. Although she provided neither figure nor measurements of the hydrothecae, she stressed that the ''trophosome of these specimens agrees entirely with those previously described'' (Millard 1980, p 138). She provided a detailed description of the coppinia, this time based on well-preserved material (Millard 1980, p 138-140). The structure of this coppinia is similar to that described by Hirohito (1995) in material assigned to A. angulata, and here considered conspecific with A. bulbosa. Schuchert's (2003) material assigned to A. rectangularis agrees with Stechow's material of A. bulbosa here studied. The specimens are similar in the size of the hydrotheca, as well as in the size of the nematocysts (2367 mm and 563 mm). However, Schuchert's figures are similar to A. rectangularis in the shape of the hydrothecae and, therefore, no decision seems possible at the moment.

Distribution
Apart from the many possible records, the species is only known with certainty from Sagami Bay (Stechow 1932;Hirohito 1995), where it was collected at 100 m (Hirohito 1995). (  Branches in distinct zigzag pattern. Hydrothecae alternate, more or less in one plane. Hydrotheca tubular, almost cylindrical in the free part; diameter softly decreasing towards basal part. Hydrotheca gently curved abcaulinarly; approximately two-thirds of adcauline wall adnate to internode. Adcauline wall convex, except the straight basal third. Abcauline wall concave, though practically straight at basal third. Sometimes, a ring of nodules visible at the base of the hydrothecae. Hydrothecal aperture circular, slightly directed upwards. Rim even, sometimes with short renovations.

Acryptolaria conferta
According to Allman (1877, p 18), the fusiform coppinia consists of a mass of closely set, flask-shaped gonothecae provided with a distal neck ending in an even circular orifice. Each gonotheca gives rise to a single ovum, subsequently lodged in an acrocyst. The coppinia lacks defensive tubes.

Remarks
According to Allman (1877, p 17-18) the hydrotheca has a distinct floor, which possibly disappears with age, being completely absent in the older hydrothecae immersed in the fascicled stem. The examination of type material has left no doubt that the hydrotheca lacks any floor, as shown in Figures 4, 17C. However, as indicated above, a ring of nodules is sometimes visible at the base of the hydrothecae. Ritchie (1911) proposed Cryptolaria conferta var. australis, differing from the nominotypical variety mainly in the growth-form and slightly smaller hydrothecae.  and Ralph (1958) considered the overlapping hydrothecae and the absence of any slight convexity at the base of the abcauline wall as the main diagnostic characters of the variety.  found those characters variable within a colony and, therefore, held that it was not justifiable to retain Ritchie's form as a separate variety or subspecies. Vervoort and Watson (2003, p 43) considered the subspecies A. c. australis and the nominotypical subspecies as conspecific. Ramil and Vervoort (1992) described a new subspecies, Acryptolaria conferta minor, as it had smaller hydrothecae than the material referred to A. conferta in their collection and did not intergrade with it (Ramil and Vervoort 1992, p 46-48). Under the synonymy of this subspecies, Ramil and Vervoort (1992, p 43) included the records of Acryptolaria gracilis by Billard (1906bBillard ( , 1906c and Patriti (1970), but not Acryptolaria gracilis by  and Ralph (1958), that would be synonyms of the valid species Acryptolaria gracilis (Allman, 1888) (see below). We believe that, without further Table IV. Morphometric data of the type material of Acryptolaria conferta (in mm).

Mean¡SD
Range n information on cnidome and a broader revision of the species records, the description of a new subspecies is premature. The material assigned to A. conferta by Hirohito (1995) has several coppiniae, either with or without protective hydrothecae, branched in rare occasions. It is likely that Hirohito's material embraces more than one species. In fact, his material probably includes a new species of Acryptolaria, since the known species described so far, with the exception of A. bulbosa and A. rectangularis, lack defensive modified hydrothecae in the coppinia, and in these species the structure of the coppinia, with the modified hydrothecae forming a canopy, and the shape of the hydrothecae are completely different.
Although Acryptolaria conferta is the most frequently recorded species of the genus, most of the records refer to sterile material and none gives information on the cnidome. Because the species has no clearly distinguishing features, we considered only the original record as representative of its geographical distribution.

Description
''A small fragment of a colony (2 cm long) with polysiphonic caulus and alternative ramifications'' (Naumov and Stepanjants 1962, p 73). Branches in a slight zigzag. Hydrothecae alternately arranged in approximately one plane. Hydrotheca horn-shaped, diameter gradually decreasing from aperture to basal part. Hydrotheca slightly curved abcaulinarly; basal half to two-thirds of adcauline wall adnate to internode. Adcauline wall slightly convex; abcauline wall slightly concave. Hydrothecal aperture circular, strongly directed upwards, forming an angle of ca 35u with longitudinal axis of branch. Rim even, sometimes with few short renovations.

Remarks
The material available consists of a microslide preparation; nematocysts are extremely difficult to find. However, putative nematocysts were seen.  Naumov and Stepanjants (1962) indicated that A. corniformis is allied with A. crassicaulis, being different in the larger dimensions of the hydrothecae and the greater length of the free adcauline part in A. corniformis.  again pointed out the similarity between A. corniformis and A. crassicaulis, but differentiated both species by the smaller hydrothecae, more curved longitudinal axis, and less pronounced difference between the diameters of the aperture and basal part in A. crassicaulis. Gravier-Bonnet (1979), however, considered both species conspecific based on the habit of the colony and measurements. After examining the type material of both species, we consider that they constitute two valid, clearly distinguishable species (see below under the discussion of A. crassicaulis).

Description
''Colony attaining a height of four inches [101.6 mm], profusely and very irregularly branched main stem'' (Allman 1888, p 41). Colony apparently composed of several stems growing tight together on a coral fragment. Longest stem ca 65 mm high.
Stem strongly polysiphonic; only distalmost branches monosiphonic. Stems branched more or less irregularly in one plane. Some basal branches becoming lower-order stems; branches up to five-order, sometimes anastomosing. Hydrothecae alternately arranged in approximately one plane. Hydrotheca tubular, cylindrical in the free part; diameter softly decreasing towards basal part. Hydrotheca curved abcaulinarly, especially at distal part; ca two-thirds of adcauline wall adnate to internode. Free part of adcauline wall distinctly convex; adnate part only slightly convex. Abcauline wall straight basally, but clearly concave at distal third. Hydrothecal aperture circular, directed upwards. Rim even, sometimes with few short renovations.

Remarks
Allman (1888, p 41) characterized this species by the ''profuseness and irregularity of its ramification, and for the great thickness of its stem and principal branches. The ultimate branches on the other hand are slender and flaccid. The hydrothecae are stout, and the exserted portion rather long. Here and there, and at uncertain intervals, slight constrictions may be noticed in the branches''. As indicated above, Acryptolaria crassicaulis has been regarded by some authors as conspecific with A. corniformis. Nevertheless, we consider that both species are valid. Acryptolaria corniformis has distinctly larger hydrothecae (cf . Table XV) and the hydrothecae are not so bent abcaulinarly as in A. crassicaulis, in which they are strongly abcaulinarly directed at distal third.
A variety of the species, Acryptolaria crassicaulis var. dimorpha, was described for the genus Cryptolaria by Ritchie (1911) from Australian waters, based on the presence of single hydrothecae. The subspecies was subsequently recorded by  for Solomon Islands. Millard (1967, p 174) properly stated that ''as the ability to produce solitary hydrothecae is a feature common to many, and possibly all, Lafoeidae, it is not necessary to retain Ritchie's variety dimorpha'', a position adopted by Gravier-Bonnet (1979, p 20).

Distribution
Off Ascension Island, 07u549200 N, 14u289200 W, 756 m, volcanic sand (Allman 1888  Stem completely polysiphonic, ca 20 mm in height. Hydrorhiza dish-shaped. Stem slightly geniculate with six first-order branches, one with incipient secondary branch. The three more basal branches also polysiphonic; polysiphony decreasing distally. Hydrothecae alternate, more or less in one plane. Hydrotheca cylindrical, diameter only slightly decreasing at most basal part, gently curved abcaulinarly, especially at distal part; from two-thirds to six-sevenths of adcauline wall adnate to internode. Adcauline wall convex. Abcauline wall concave at distal half, though practically straight at basal half. Hydrothecal aperture circular, parallel to longitudinal axis of branch or slightly directed upwards or downwards. Rim even, usually with numerous renovations (four to 90) which curve hydrothecal aperture downwards; the most distal hydrothecae with few short renovations. Hydrothecal perisarc longitudinally striated.

Remarks
The material examined lacked coenosarc, and it was almost impossible to find nematocysts; a single putative nematocyst was observed.
As noted by Allman (1888, p 40), the rigid habit, the ramification in one plane and, especially, the long curved hydrothecae make the species quite distinctive from other species of Acryptolaria.
The study of the type material of Cryptolaria flabellum has confirmed that it belongs to the genus Acryptolaria, as was already indicated by Fraser (1944) and Vervoort (1968), who based their opinions particularly on the alternate pattern of the hydrothecal arrangement. Vervoort (1972) studied material of a species of Cryptolarella from deep waters of the tropical eastern Pacific, west of Peru, and considered it conspecific with Allman's Cryptolaria flabellum-this fact supported his decision in transferring Allman's species to the genus Cryptolarella. Although the material studied by Vervoort clearly belongs to the genus Cryptolarella, as can be concluded by the morphology of its gonothecae (cf. Marques et al. 2005), Allman's species is undoubtedly a species of Acryptolaria. Naumov (1960) described rare material sampled in the Bering and Okhotsk Seas as Cryptolaria flabellum, but Vervoort (1972) regarded it as belonging to a new species, distinct from Cryptolarella flabellum, because of the shorter hydrothecal length, the more dense sets of hydrothecae, the lack of the initial arrangement in slightly displaced pairs, and the extremely high number of distal hydrothecal renovations. According to Marques et al. (2005), Naumov's material seems similar to the type species of Cryptolarella abyssicola, except for the presence of numerous renovations in the older (5basal) hydrothecae. This and the fact that Naumov's material was not sufficiently described (infertile material) makes it impossible to evaluate its specific status, although we consider plausible that it belongs, indeed, to Acryptolaria flabellum.
Acryptolaria gracilis (Allman, 1888) (Figures 8, 15F   Stem fragments irregularly branched in several planes. Fragment 18 mm long with secondary branches. Accessory tubes almost reaching the most distal parts of branches.
Hydrothecae alternate, more or less one plane. Hydrotheca tubular, cylindrical at distal half, diameter approximately constant from hydrothecal aperture to the middle of hydrothecal length, then slightly decreasing up to become more or less constant, forming a kind of bottleneck at the most basal part. Hydrotheca gently curved abcaulinarly; approximately three-quarters of adcauline wall adnate to internode. Adcauline wall mostly convex, but concave at basal fourth. Abcauline wall straight or slightly convex at basal half, concave at distal half. Hydrothecal aperture circular, directed upwards, forming an angle of ca 45u with longitudinal axis of branch. Rim even, sometimes with up to five short renovations.

Remarks
There is some confusion regarding the naming of the type material of this species. Allman (1888) did not specify holotypes in his paper and part of the material studied by him is now deposited in the collections of the ZMUC (labelled as ''cotype'') and the BMNH. The International Commission on Zoological Nomenclature (1999) considered that ''cotype'' is a term not recognized by the Code, formerly used for either syntype or paratype, and should not be used anymore in zoological nomenclature (Recommendation 73E); in the case of colonial material, cotype was also used in the sense of ''part of the holotype material''. On the other hand, the BMNH material was referred as ''holotype'' by Vervoort and Watson (2003), who also referred to a ''paratype'' present in the collection of the Nationaal Natuurhistorisch Museum (RMNH). Although it is possible to consider the existence of split holotypes (with the same collection data of the original description of the species), ''paratypes'' may not be considered as valid, as there was no such designation by Allman (1888). In the case that the original material is regarded as being ''syntypes'', a lectotype should be selected.  Vervoort and Watson (2003), after examining part of the supposed ''holotype'' in the Natural History Museum (88.11.13.31), found this species to be quite near to Acryptolaria conferta, noting that it was extremely difficult to define characters by which the two could be distinguished. Finally, they pointed out that both species may well be conspecific. Actually, the association of nominal records of A. gracilis and A. conferta is long known in the literature, such as the synonymy of the records of A. gracilis by Billard (1906bBillard ( , 1906c and Patriti (1970) (cf. Ramil and Vervoort 1992, p 43; see above). In our opinion, however, both species are clearly distinguishable. Although the hydrothecae are of similar size (cf .  Table XV), in A. conferta the free portion of the adcauline hydrothecal wall is distinctly longer, its hydrothecae lack the basal bottleneck (the most distinctive character for A. gracilis), and the branches form a clear zigzag. Finally, A. gracilis has larger nematocysts (19.5-2167.5-9 mm in A. conferta).

Distribution
Off East Cape, New Zealand, 37u349 S, 179u229 E, 1260 m, mud.  (Allman 1877, p 19). Presently, type material consisting of numerous small fragments up to 15 mm in length. Fragment 15 mm long belonging to basal part of stem, completely polysiphonic, slightly geniculate and with a basally flat hydrorhiza probably indicating attachment to hard substratum. Hydrothecae alternately arranged almost in one plane, forming an obtuse angle. Hydrotheca tubular, cylindrical in the free portion, then diameter slightly decreasing to basal part. Hydrotheca strongly curved abcaulinarly; approximately half of adcauline wall adnate to internode. Free portion of adcauline wall practically straight, adnate part slightly convex. Abcauline wall concave. Hydrothecal aperture circular, directed upwards, forming an angle of ca 60u with longitudinal axis of branch. Rim even, sometimes with few short renovations. Coppinia not observed.

Remarks
All material studied by us and assigned to A. longitheca seems to belong to the same colony that corresponds to the holotype of the species. The jar in which the type material of A. longitheca is preserved included a small, unlabelled vial with colonies of A. abies. Allman (1877, p 19) indicated that ''Cryptolaria longitheca is a far less profusely branched species than C. conferta and from this species it further differs in the pinnate disposition of its branches and in being a stronger form with larger hydrothecae. The hydrothecae, moreover, … are cylindrical throughout, presenting no diminution of their diameter towards the base as in C. conferta. They appear also to pass continuously into the tubes of the hydrocaulus, no distinct floor being apparent in the hydrothecae of any part of the specimen''. Calder and Vervoort (1998) examined type material of A. longitheca and A. crassicaulis. They found nothing to warrant their separation and, consequently, considered A. crassicaulis conspecific with A. longitheca. Although we recognize similarities between the two species, we regard them as valid (see below). The material studied by these authors is similar to A. crassicaulis, because of the shape of the hydrothecae, softly curved abcaulinarly, and the much larger adnate proportion of the adcauline wall. In A. longitheca the hydrotheca is strongly curved abcaulinarly and approximately half of the adcauline wall is adnate to the internode. Unfortunately, no information concerning the cnidome of the material studied by those authors is available, and it is not possible, at present, to evaluate their identifications based on nematocyst characters.
Acryptolaria longitheca belongs to a group of Acryptolaria species characterized by large hydrothecae with a long free portion of the adcauline wall, and includes A. corniformis, A. crassicaulis, and A. operculata (cf . Table XV). However, it is clearly distinguishable from these species. Acryptolaria operculata has much larger general dimensions (the differences between them will be discussed below, when dealing with that species). Acryptolaria corniformis is also a larger species with hydrothecae slightly bent abcaulinarly, forming an angle of ca 35u with the long axis of branch, the free portion of the adcauline wall is distinctly larger (680-930 mm), and the larger size-group nematocysts are distinctly smaller (10-1563-5 mm). Acryptolaria longitheca has strong affinities with A. crassicaulis, though it has slightly larger hydrothecae, with a larger proportion of the adcauline wall adnate to the internode and with distinctly smaller larger nematocysts (14-16.564.5-5.5 mm).
Clarke (1879) recorded Cryptolaria longitheca from off Tortugas. We considered Clarke's record as valid, because it completely agrees with Allman's species in shape and size of the hydrothecae, and in the free proportion of the adcauline wall. Clarke found coppinia and described them as ''clustered upon the upper portions of the stem of one of the finest specimens were a number of peculiar bodies … They are polygonal in form, largest at the distal end, tapering to the base, crowded so closely together that the walls of adjoining bodies are in contact throughout their length, and are provided with a small tubular orifice arising from the centre of the distal end; at the base they are connected by branching stolons''.

Description
Palmate colony, ca 90 mm in height, composed of two main stems, one of them basally giving rise to a conspicuous secondary stem. Stems strongly branched, giving rise to numerous lower-order branches (up to fourth order). Stems almost completely polysiphonic, branches polysiphonic up to distalmost ends, though in most distal parts usually as a single tube. Colonies alternately branched in one plane, though sometimes branches arising either in decussate pairs or alternately in several planes. Branches frequently anastomosed.
Hydrothecae alternate, in one plane. Hydrotheca tubular, almost completely cylindrical; diameter slightly decreasing at basal part. Hydrotheca gently curved abcaulinarly, adcauline wall approximately four-fifths adnate to internode. Adcauline wall convex; abcauline wall mostly concave, but slightly convex at basal part. Hydrothecal aperture circular, directed upwards, forming an angle of 40-60u with long axis of branches. Rim even, usually one to two renovations in most distal hydrothecae; renovations more numerous at basal and polysiphonic parts.

Remarks
Acryptolaria minima is well characterized by the shape and size of the hydrothecae. In fact, it is the species with the smallest hydrothecae amongst the known species of Acryptolaria (cf .  Table XV). By contrast, apart from A. tortugasensis (see below), it has the largest nematocysts. Acryptolaria minima is also characterized by having hydrothecae almost completely adnate to the branches (only one-fifth of the adcauline wall is free). The wellmarked characters make it possible to corroborate some of the identifications made in the literature, such as that of Ralph (1958) for a specimen from Cook Strait, New Zealand.

Distribution
Acryptolaria minima seems to be a shelf species, having been found at depths between 20  and 180 m (Ralph 1958). Presently, it is known only from New Zealand waters, where it has been reported from off Cape North and Cape Maria van Diemen  and from Cook Strait (Ralph 1958 Figure 13A-C.

Description
''Small infertile colony of five small stems, two branched … Tallest stem 9 mm high, broken off at tip, 3.5 mm wide at base; stems branched once in one plane; branched stems arising from a small matted plug of stolons, simple stems from junction of stolonal tubes; taller stems with up to four polysiphonic tubes intergrown and rather contorted proximally, becoming linear distally; polysiphonic tubes running almost to top of stems; stems lightly fascicled; branches given off beside a hydrotheca'' (Watson 2003, p 163).
Hydrothecae alternately arranged in one plane. Hydrotheca tubular, long and thin, diameter slightly larger at free part, clearly decreasing at basal part. Hydrotheca strongly curved abcaulinarly, ca two-thirds adnate to internode. Adcauline wall convex; abcauline wall markedly concave. Hydrothecal aperture circular, upwards and abcaulinarly directed, forming an angle of 60-90u with long axis of branches. Rim even, with few renovations. Coppinia not observed.

Remarks
According to Watson (2003, p 164), ''their smaller dimensions and sparsely branched habit matches no other known species of Acryptolaria''. She considered this species similar to A. patagonica (see remarks for this name under the description of A. operculata below) in habit, though considerably smaller. Watson also indicated that ''The slightly everted margin of the hydrotheca resembles A. conferta minor Ramil and Vervoort, 1992 but it is smaller and the hydrothecae are frontally directed in contrast to those of A. conferta minor which, according to these authors, lie in the plane of ramification of the branches'' (Watson 2003, p 164). Acryptolaria minuta is similar to A. minima in the size of the hydrothecae (cf . Table XV), though in the former the hydrothecae are larger and thinner, strongly curved abcaulinarly, and with a much larger proportion of their adcauline wall free. Moreover, the nematocysts are much smaller (25-2769-10 mm in A. minima).
The material described as Acryptolaria conferta var. australis by Ritchie (1911) and  (the latter only the material from Sta. 96) might belong to this species.

Distribution
Acryptolaria minuta is known only from the Macquarie Island area (52u59.49-53u029 S, 159u599-159u58.29 E), where it was collected at a depth of 1422 m on dead primnoid gorgonian.
Hydrothecae alternate, in two planes, making an obtuse angle. Hydrotheca tubular, fusiform; maximum diameter approximately in the middle of hydrothecal length, decreasing distinctly towards basal part. Hydrotheca curved abcaulinarly, half to twothirds of adcauline wall adnate to internode. Adcauline wall convex; abcauline wall concave. Hydrothecal aperture circular, directed upwards, forming an angle of 55-75u with long axis of branch. Rim even, with few short renovations.

Remarks
Acryptolaria operculata is characterized by the size and arrangement of the hydrothecae. It is the species with the largest hydrothecae (cf. Table XV) and it is also characterized by the hydrothecal arrangement in two planes, making an obtuse angle, though this feature is also present in other species of the genus (e.g. A. tortugasensis, see below).  also characterized this species by the presence of a one-flap operculum, pointing out the particular position of this species amongst the Lafoeidae. However, after studying many colonies of Acryptolaria, we have observed that the presence of that circular flap, in spite of being irregular, is quite common amongst species of Acryptolaria, being also common in species of another lafoeid genus, Grammaria. That valve should not be considered as an opercular apparatus in the sense of that present in other families of hydroids (e.g. Sertulariidae or Campanulinidae). By contrast, it seems to constitute a sort of deciduous structure to temporally protect the polyps when they are in a state of inactivity or regeneration, as has already postulated for other groups such as Staurotheca (cf. Peñ a Cantero et al. 1997).
In his PhD dissertation on hydrozoans from the Patagonian shelf, El Beshbeeshy (1991) included the description of 22 new species, as well as many re-descriptions and some nomenclatural considerations. Many subsequent papers (e.g. Peñ a Cantero et al. 1997;Peñ a Cantero and Marques 1999;Vervoort and Watson 2003;Watson 2003) have used the names employed by El Beshbeeshy (1991). However, according to the International Code of Zoological Nomenclature (1999), a thesis is not to be taken as published for the purposes of zoological nomenclature, and therefore the specific names proposed by El Beshbeeshy (1991) should be considered nomina nuda. According to the Code ''A nomen nudum is not an available name, and therefore the same name may be made available later for the same or a different concept; in such a case it would take authorship and date [Arts. 50, 21] from that act of establishment, not from any earlier publication as a nomen nudum''.
Indeed, many nominal species of El Beshbeeshy were subsequently described by other authors and, in this case, these should become the authors of the valid names. Nevertheless, Article 50.1.1. of the Code establishes that ''however, if it is clear from the contents that some person other than an author of the work is alone responsible both for the name or act and for satisfying the criteria of availability other than actual publication, then that other person is the author of the name or act. If the identity of that other person is not explicit in the work itself, then the author is deemed to be the person who publishes the work''. El Beshbeeshy (1991) described A. patagonica as a new species; the species was again reported by Watson (2003) and Vervoort and Watson (2003), and both papers clearly gave the authority of the species to El Beshbeeshy. Therefore, the species is herein assumed as ' 'Acryptolaria patagonica El Beshbeeshy in Watson, 2003''. In any case, the species nominally described as A. patagonica would be characterized by its large hydrothecae and the arrangement of the hydrothecae in two planes, forming an obtuse angle. According to El Beshbeeshy (1991), A. patagonica differs from A. operculata by the absence of operculum, the shorter free part of the adcauline wall (672-858 mm), and the larger diameter at hydrothecal base (232-324 mm). Nevertheless, we have re-examined some slides of the material described by El Beshbeeshy and there are no differences between it and the type of A. operculata. Additionally, both have similar dimensions of their nematocysts (13.5-1464.5-5 mm and 863-3.5 mm in A. patagonica). Therefore, we consider A. patagonica conspecific with A. operculata. Vervoort and Watson (2003) also assigned material from New Zealand to A. patagonica and, again, the description fully agrees with the type material of A. operculata. Vervoort and Watson (2003) indicated that the most significant character in A. patagonica is that the conspicuous hydrothecae are not in one plane, but point obliquely frontally. As we have demonstrated before, this feature is also found in the type material of A. operculata, and it is clearly described by  in her original description.

Distribution
Acryptolaria operculata seems to be a shelf and bathyal species, having been found at depths between 98 (El Beshbeeshy 1991) and 1422 m (Watson 2003). Watson (2003) found it on primnoid gorgonians. It is known to be distributed in sub-Antarctic waters, from the Patagonian shelf El Beshbeeshy 1991), New Zealand (Vervoort and Watson 2003), and Macquarie Island (Watson 2003
Stem strongly polysiphonic, quite straight basally, giving rise to branches alternately arranged in one plane, sometimes in two planes making an obtuse angle. Distal part of Table XIII. Morphometric data of the type material of Acryptolaria pulchella (in mm).

Mean¡SD
Range n  colony with indefinite branching pattern; some areas with alternate branching, others with irregular arrangement. Basal part of stem with branches originating at ca 45u, but thereafter curving upwards.
Hydrothecae alternate, in one plane. Hydrotheca tubular, almost completely cylindrical; diameter only slightly decreasing at basal part. Hydrotheca gently curved abcaulinarly; four-fifths to three-quarters of adcauline wall adnate to internode. Adcauline wall entirely convex; abcauline wall concave. Hydrothecal aperture circular, directed upwards, forming an angle of 50-70u with long axis of branch. Rim even, with few short renovations.
Coppiniae without defensive tubes. Gonothecae fusiform, bottle-shaped, firmly aggregated. Diameter strongly decreasing at distal part forming a short distal neck bearing a circular aperture.

Remarks
''The very regularly pinnate disposition of the ramuli, and the absolutely distichous and regular disposition of the hydrothecae, give to this species an aspect of considerable elegance. It is a strong growing form with rather close-set, stout, and short hydrothecae'' (Allman 1888, p 41). Although Allman (1888, p 41) indicated that the ''gonosome [was] not known'', the type material examined has coppinia and afforded us an opportunity to describe the reproductive structures. Acryptolaria pulchella is characterized by the relatively short and wide hydrothecae, adnate to the branches by most of its adcauline wall, and by the large nematocysts. Description ''The material consists of a part of a colony 0.7 cm high and lacking both basal and distal portions. Fascicling tubes are few and limited to the proximal end of the stem. Branches are irregular, or perhaps opposite, and polysiphonic for a short distance beyond their origin. The hydrothecae are alternate, long and tubular, the proximal halves erect and adnate, the distal diverging at a right angle. The margin is smooth with several reduplications and circular aperture. A well-marked fold occurs in the lower wall of the hydrotheca at the point of divergence. Gonosomes absent'' (Jarvis 1922, p 335-336).

Remarks
Unfortunately, it has not been possible to examine type material of this species. According to Vervoort and Watson (2003), the location of the type material of Jarvis' species is unknown.
The branches represented in Jarvis ' (1922) figures have a zigzag arrangement. According to Vervoort (1966, as A. angulata) ''The stems are erect and largely monosiphonic; there are some accessory tubes at the basal parts of some of the hydrocauli and some short hydrorhizal fibres''.
As already noted, the species A. angulata, A. bulbosa, and A. rectangularis share similar hydrothecal morphology and may be phylogenetically closely related species. In fact, Acryptolaria rectangularis has been considered conspecific with A. angulata by several authors (e.g. Hirohito 1995; Vervoort and Watson 2003). Vervoort and Watson (2003, p 41), for example, indicated that they studied ''great amount of material of this species [A. angulata] from all over its large area of distribution … has convinced us that Cryptolaria angulata Bale, 1914 andCryptolaria rectangularis Jarvis, 1922 are inseparable''.
We found, however, that the three species are different, although sharing similar hydrothecae. Acryptolaria angulata is characterized by the presence of the adcauline boss projecting into the hydrothecal cavity.  made it clear that A. rectangularis differs from A. angulata in the straightness of the upper wall of the hydrotheca, in its divergent portion, and in the absence of an internal thickening opposite the fold. Gravier-Bonnet (1979, p 18) followed the opinion of  and Millard (1967 that this species is clearly distinguishable from Bale's species by the absence of the adcauline thickening of the hydrothecae. Schuchert (2003, p 155) also considered A. rectangularis and A. angulata as ''separate species because both morphotypes were found not very far apart''; but he acknowledges, however, ''the possibility that both could be only forms belonging to the same species''.
The presence of the adcauline boss projecting into the hydrothecal cavity is clearly a distinct character of A. angulata. On the other hand, the absence of this boss makes A. bulbosa and A. rectangularis morphologically similar. However, the shape of the hydrotheca of these two species seems to be slightly different. Stechow (1932) indicated that A. bulbosa differs from A. rectangularis mainly in the stronger bend of the hydrotheca, the smaller diameter of the hydrothecal aperture and the shorter free portion of the hydrotheca (cf .  Table XV). These differences, though small, prevent us from considering the two species to be conspecific. In addition, the hydrotheca of A. bulbosa is similar to that of A. angulata, it distinctly bends upwards after the adcauline wall becomes free. By contrast, A. rectangularis has the free portion of the adcauline wall straight, the diameter of the hydrothecal aperture larger, and the free portion of the hydrothecal wall longer than in the other two species.
Possible misidentifications of A. rectangularis, especially by Millard (1967, Gravier-Bonnet (1979), and Schuchert (2003), are listed under the discussion of A. bulbosa. On the other hand, Vervoort (1966) assigned to A. angulata material belonging to Jarvis' species. The material shares with A. rectangularis the shape and size of the hydrothecae and the absence of the adcauline notch so characteristic for Bale's species.

Description
One colony composed of a distally truncated, erect stem ca 60 mm in height, completely polysiphonic except for distal ends of branches. Main stem giving rise to lateral branches in approximately one plane, usually in an opposite or sub-opposite pattern, but with irregularities. Branches directed upwards with an angle of ca 60u or larger. Lateral branches polysiphonic, up to third-order, usually alternate or sub-opposite in one plane.
Hydrothecae alternately arranged in two planes, forming a wide angle, sometimes close to 90u. Hydrotheca cylindrical, straight at adnate part, strongly curved outwards when it becomes free. Free part of adcauline hydrothecal wall convex. Abcauline wall straight basally, becoming concave at distal third with a marked inflexion point associated to a conspicuous internal cusp. Hydrothecal aperture circular, almost parallel to long axis of branch. Rim even, though slightly everted, with one or two short renovations.

Remarks
Although this species has been considered conspecific with A. conferta by several authors (e.g . Fraser 1944;Vervoort 1968;Calder 1991), it can be distinguished from all other species. Acryptolaria tortugasensis is unique in the internal cusp present in the abcauline hydrothecal wall. It is also characterized by the obtuse angle, close to 90u, formed by the        alternate hydrothecae. It is also the Acrytolaria species with the largest nematocysts (cf .  Table XV).

Distribution
Acryptolaria tortugasensis is only known from Tortugas, Florida, USA.
Key to the known species of Acryptolaria (A. andersoni excluded)