Global Diversity and Phylogeny of Pelagic Shrimps of the Former Genera Sergestes and Sergia (Crustacea, Dendrobranchiata, Sergestidae), with Definition of Eight New Genera

We revise the global diversity of the former genera Sergia and Sergestes which include 71 valid species. The revision is based on examination of more than 37,000 specimens from collections in the Natural History Museum of Denmark and the Museum of Natural History, Paris. We used 72 morphological characters (61 binary, 11 multistate) and Sicyonella antennata as an outgroup for cladistic analysis. There is no support for the genera Sergia and Sergestes as they have been defined until now. We define and diagnose eight genera of the former genus Sergia (Sergia and new genera Gardinerosergia, Phorcosergia, Prehensilosergia, Robustosergia, Scintillosergia, Challengerosergia, and Lucensosergia) and seven genera of the former genus Sergestes (Sergestes, Deosergestes, Eusergestes, Allosergestes, Parasergestes, Neosergestes, and a new genus Cornutosergestes). An identification key is presented for all genera of the family Sergestidae. The phylogeny of Sergestidae is mainly based on three categories of characters related to: (1) general decapod morphology, (2) male copulatory organs, and (3) photophores. Only simultaneous use of all three character types resulted in a resolved tree with minimal Bootstrap support 75 for each clade. Most genera are interzonal mesopelagic migrants, some are benthopelagic (Scintillosergia, Lucensosergia), bathypelagic (Sergia), or epipelagic (Cornutosergestes). Within each of meso- and benthopelagic genera there is one species with panoceanic distribution, while most species ranges are restricted to a single ocean. The genera demonstrate two different strategies expressed both in morphology and behavior: protective (Eusergestes, Sergestes, Cornutosergestes, Prehensilosergia, Scintillosergia, Lucensosergia, Challengerosergia, Gardinerosergia, Robustosergia, Phorcosergia, Sergia) and offensive (Neosergestes, Parasergestes, Allosergestes, Deosergestes).


Introduction
The decapod suborder Dendrobranchiata (Crustacea, Malacostraca) includes shrimps that have an important role both ecologically and economically in marine and estuarine ecosystems. The approximately 500 extant species range from shallow waters in the tropics to depths of about 1000 m on the continental slopes [1]. Species of the Sergestidae are among the most common in many ecosystems [2] - [3] and important objects of fisheries in some areas, such as Sergestes lucens in Japan (Fig. 1a). Despite their importance, the sergestids are still poorly understood with regard to higher level classification and phylogenetic relationships.
The two genera, Sergestes H. Milne-Edwards, 1830 (Fig 1b) and Sergia Stimpson, 1860 (Fig. 1c), together most certainly form a monophyletic group based on a number of synapomorphies such as the presence of an organ of Pesta and dermal photophores [4] - [8] and comprise 71 species, i.e. two-thirds of all known recent sergestids. The taxonomic status of Sergestes and Sergia and the phylogenetic relationship of their constituent species are the object of the present paper. The taxonomic history of Sergestes and Sergia goes back to the first half of the nineteenth century when H. Milne-Edwards [9] and Stimpson [10] described Sergestes and Sergia, respectively. Among later researchers, the most productive and important was H. J. H. Hansen [4], [11][12][13][14][15][16][17], who critically reviewed the existing knowledge and generated much new data on sergestid taxonomy and classification. Hansen [4], [11] synonymized Sergia with Sergestes, which he later subdivided into two species groups [12], [14]. Yaldwyn [5] synthesised the available information about the taxonomy and morphology of Sergestes (sensu Hansen [4], [11]), dividing the genus into two subgenera, Sergestes and Sergia, and nine species groups (six in Sergestes and three in Sergia). Later researchers mostly followed Yaldwyn's classification at least until the mid-1970s, when Omori [6] reviewed the differences between Yaldwyn's subgenera Sergestes and Sergia both with respect to morphology and ontogeny and raised their taxonomic status to the generic level. Most recent authors have followed the separation of the two genera (for example, [7] - [8], [18]).
As a result of the work of Vereshchaka [7] - [8], [19], which involved the checking of all available type material and documenting of intra-and interspecific variation of many characters, six species were synonymized and 13 new species were recognized, yielding a total number of 71 species in Sergestes and Sergia.
During this revisionary work it was noted that both Sergia and Sergestes contained taxonomic groups that appeared sufficiently well-defined to be treated as genera, but such a step was at the time postponed until it had been tested by a more comprehensive phylogenetic analysis. Recently, Judkins and Kensley [18] erected five new genera of Sergestes replacing Yaldwyn's species groups. They provided diagnoses for the new genera mostly based on the characters from Yaldwyn's work but provided no new evidence for the support of the new taxa.
Previous attempts to classify sergestids above the species level have focused on various subsets of characters (e.g., male copulatory organs or photophores) which have yielded conflicting phylogenetic and classificatory signals [7] - [8].
In this paper we present a phylogenetic approach based upon extensive studies of most available specimens and using all character systems (general external morphology, copulatory organs such as the petasma and clasping organ in males, and photophores). We use the resulting phylogeny to present a new classification, and to explore morphological, ecological, and biogeographical patterns within the clades.

Material examined
This study is primarily based on material in the Natural History Museum of Denmark (former 'Zoological Museum, University of Copenhagen') and Museum National d'Histoire Naturelle (Paris), which hold the richest collections of sergestids in the world. In Copenhagen the primary contribution of sergestids came from the expeditions ''Dana-1'' and ''Dana-2'' (Table 1). In Paris the primary contribution came from ''Caride I-V'' and ''Cyclone 3-6'', specimen numbers for studied samples have not established. Much of these two museums' sergestid material was studied by Vereshchaka [7] - [8], and more than 28,000 specimens of Sergestes and 9,000 specimens of Sergia were examined; much information about the examined material such as distribution and information about type specimens can also be found in Vereshchaka's papers [7] - [8].

Morphology and Characters
The morphological information provided by Vereshchaka [7] - [8] forms the basis of the phylogenetic analysis of Sergestes and Sergia presented here. Seventy-one species were included in the analysis, 36 belonging to the former Sergestes, 35 to the former Sergia (Table 2). Seventy-two morphological characters were identified (61 binary, 11 multistate). The characters used in this study fall into three main categories: 24 general external characters, 26 characters related to male copulatory organs, and 22 characters related to photophore structures.
1. General external characters relate to external morphology such as the feeding/catching limbs (mainly maxilliped III, pereopods I-III). In general, external morphology shows little diversity. 2. Male copulatory organs include the petasma and the clasping organ (modified part of Antenna I). The petasma is a particularly complicated structure which exhibits much variation; for example, the petasma is sometimes very different in species that are otherwise very similar (closely related), or, vice versa, very similar in species that are otherwise very different (distantly related). 3. Photophores. These organs are present in most species of the genus Sergia. In photophores, a lens may be absent or present. A challenge in using photophores is that they fade away in alcohol preserved material. Reliable information can often only be obtained for flat and transparent body parts (i.e. scaphocerites and uropods).
Petasmas and photophores are the two classical organ systems used in the classification and systematics of the Sergestidae. However, in many cases photophore arrangements and petasma structure are incongruent. Species having similar photophore arrangements may have different petasmas and vice versa. This will be explored further below. Detailed information on the morphology underlying the characters used in this work can be found in [7] - [8], but some of the key characters are illustrated here (Figs 2-7). Our terminology follows [7] - [8].

Anatomical Abbreviations
We label most general anatomical characters directly on the figures but structures relating to the petasma, are abbreviated as follows: LA 2 lobus armatus LAc 2 lobus accessorius LC 2 lobus connectens LI 2 lobus inermis LT 2 lobus terminalis PU 2 processus uncifer PV 2 processus ventralis Description of characters and their states is presented in Table 3.

Phylogenetic Analysis, Outgroup, and Character Optimization
Data were handled and analyzed using a combination of a number of programs using maximum parsimony: Winclada/ Nona, NDE (Nexus Data Editor), TNT, and Mesquite [20] - [22]. The trees on which the classificatory changes (e.g., erection of six  Table 2. Names of old and new taxa within Sergestes sensu Hansen (1903;1919) [4], [13].
Subgenera sensu Burkenroad (1937Burkenroad ( , 1945 [23], [43] and Yaldwyn (1957) [5], equivalent to genera sensu Omori (1974) [6] Species groups sensu Yaldwyn (1957) [5] Species groups sensu 2009)  new genera) and evolutionary consideration (e.g., zoogeography) are built were generated in TNT using the 'traditional search' options. The search parameters were set to the following: memory set to hold 1.000.000 trees; 1000 replicates with tree bisectionreconnection (TBR) branches swapping and saving 1000 trees per replicate; zero-length branches collapsed; suboptimal trees were set to be filtered out. Sicyonella antennata (Sergestidae) was used as the outgroup. This species is clearly outside the Sergia/Sergestes species complex, and, following [23], there is some evidence that its morphology may be close to a common sergestid ancestor. It is common opinion (to be tested phylogenetically) that pereopods have been gradually reduced within Sergestidae possibly as an adaption to a pelagic life style. Sicyonella antennata (outgroup), which occurs in near-bottom water layers [23] - [24] where many groups of shrimps are known to have evolved [25], has, with other species in the genus, the least reduced pereopods within the family, e.g., with chelae on pereopod 1 and dactyli on pereopods 4 and 5. Species of this genus also have well-developed gills (branchiae) and no luminous organs.
Characters were mapped on trees using TNT's character mapping functions. Some of these characters (synapomorphies) are listed in Table 4 and shown on the strict consensus tree in Fig. 8.
In general, we preferred character optimizations that favored primary homology (ACCTRAN).
Bremer support was calculated with the use of the program TNT, algorithm TBR, with the following settings: for all suboptimal trees, trees searched with a score of up to 2-15 worse than best (in 2-15 searches, each one worse than previous), saving up to 1000 trees.
Bootstrap support was calculated with the use of the program TNT, standard (sample with replacement), with the following settings: 10000 replicates, traditional search, groups collapse below 75.

Nomenclatural Acts
The electronic edition of this article conforms to the requirements of the amended International Code of Zoological Nomenclature, and hence the new names contained herein are available under that Code from the electronic edition of this article. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix ''http://zoobank.org/''. The LSID for this publication is:  Burkenroad (1937Burkenroad ( , 1945 [23], [43] and Yaldwyn (1957) [5], equivalent to genera sensu Omori (1974) [6] Species groups sensu Yaldwyn (1957) [5] Species groups sensu 2009) [7] - [8] Genus sensu Junkins and Kensley (2008) [18] Genus 1573AF28-5DD4-47ED-AACD-2C2DDCB47E02. The electronic edition of this work was published in a journal with an ISSN, and has been archived and is available from the following digital repositories: PubMed Central, LOCKSS.

The Clades
All equally weighted parsimony analyses of the complete data set in TNT (traditional search) resulted in 8 equally short trees. The strict consensus tree is rather resolved, especially the deeper nodes, so classificatory conclusions have been based on this tree ( Fig. 8). Further we consider only the clades corresponding to the genera, supported by synapomorphies, and having Bootstrap support 75 or more. There is no support for Sergia and Sergestes genera as they have been defined until now. Sergia and Sergestes are therefore both redefined with a less inclusive content as are a number of smaller genera, some new and some proposed by [18].
All characters fall into one of three groups: (1) general decapod anatomy, (2) male sexual characters (clasping organ, petasma), and (3) photophore patterns. Table 4 shows that the clades are supported by synapomorphies relating to both sexual and nonsexual characters in different proportions for various clades.
Clade 2 is supported by 3 synapomorphies related to sexual structures and includes the only species Sergestes atlanticus. Our phylogenetic analysis supports generic status of this group (isolated species in [8]).
Clade 3 is supported by one non-sexual synapomorphy and 2 synapomorphies related to sexual structures and includes the former ''Sergestes cornutus'' species group (erected by Vereshchaka [8]). Bootstrap support of this group is remarkably high (92). The clade includes two species. On the basis of phylogenetic analysis we raise the status of this species group to generic level. In order to maintain the continuity of the systematics of Sergestes and Sergia we name the new genus Cornutosergestes.
Clade 4 is supported by one non-sexual synapomorphy and one synapomorphy related to sexual structures and corresponds to the former ''Sergestes edwardsi'' species group [5], [8] and the genus Neosergestes [18]. The clade includes six species. Our phylogenetic analysis supports the genus Neosergestes which was erected by Judkins and Kensley [18].
Clade 5 is supported by two general morphology synapomorphies and one synapomorphy related to sexual structures. Bootstrap support of this group is very high (93). The clade includes seven species and corresponds to the former ''Sergestes vigilax'' species group [5], [8] and the genus Parasergestes [18]. Our phylogenetic analysis supports the genus Parasergestes erected by Judkins and Kensley [18].
Clade 6 is supported by one general morphology synapomorphy and one synapomorphy related to sexual structures and corresponds to the former ''Sergestes corniculum'' species group [5], [8] and the genus Deosergestes [18]. The clade includes 8 species. Our phylogenetic analysis supports the erection of the genus Deosergestes by Judkins and Kensley [18].
Clade 7 is supported by 2 general morphology synapomorphies and 2 synapomorphies related to sexual structures and corresponds to the former ''Sergestes sargassi'' species group [5], [8] and the genus Allosergestes [18]. This is the best supported clade (Bootstrap support 97, Bremer support 4). The clade includes eight species. Our phylogenetic analysis supports the genus Allosergestes which was erected by Judkins and Kensley [18].
Clade 8 is supported by one synapomorphy related to general morphology. This clade includes 4 species belonging to the former ''Sergia japonica'' species group along with the isolated species Sergia tenuiremis and Sergia inoa [7]. We recognised this clade as a separate genus Sergia. Clade 9 is supported by 2 synapomorphies related to sexual structures and corresponds to the former ''Sergia robusta'' species group [7]. The clade includes four species and is recognised as a new genus Robustosergia.
Clade 10 is supported by one sexual synapomorphy and 6 synapomorphies related to photophore characters. The clade corresponds to the former ''Sergestes phorca'' species groups [7], includes 9 species, and is recognised as a new genus Phorcosergia.
Clade 11 is supported by one synapomorphy related to general morphology and 2 synapomorphies related to photophore characters. The Bremer support is significant for this clade. The clade corresponds to the former ''Sergia gardineri'' species group [7], includes five species, and is recognised as a new genus Gardinerosergia.
Clade 12 is supported by one sexual synapomorphy and 3 photophore-related synapomorphies and includes a single species, Sergia prehensilis. Vereshchaka [8] placed Sergia prehensilis and Sergia scintillans together in a ''Sergia prehensilis'' species group. Our phylogenetic analysis splits this group into two parts with Sergia prehensilis placed in a separate monotypic genus. Clade 12 is recognised as a new genus Prehensilosergia.
Clade 13 is supported by 2 synapomorphies related to sexual structures and includes the single species Sergia scintillans, which is recognised now as a new monotypic genus Scintillosergia.
Clade 14 is supported by 2 sexual synapomorphies and one photophore-related synapomorphy and corresponds to the former ''Sergia lucens'' species group [7]. The clade includes four species and is now recognised as a new genus Lucensosergia.
Eusergestes Judkins, Kensley, 2008 Diagnosis: Integument firm, frontal margin of rostrum oblique, supraorbital tooth present, hepatic protrusion forming tooth. First segment of A I elongate,.1.5 times as long as 3d segment; distoventral end of 3rd segment in males without process; distal tooth of scaphocerite reaching or overreaching end of blade; maxilliped I developed, 3-4-segmented, maxilliped III ,2.0 times as long as Cp, dimorphic sexually, dactyl subdivided into 6 specialized subsegments; pereopods I-II ischia without strong movable spines and distally curved tooth; pereopod II with merus lacking protrusion and chela bearing equal fingers, without long setae; chela of pereopod III without strong curved spines and without long setae on propodus, fixed finger subequal to dactyl; pereopods IV-V present, 6-segmented; pereopod V setose along one margin; uropodal exopod with outer spine, proximal segment not setose along outer margin. Male clasping organ: developed, without serrated bristles, tubercle present. Petasma: LA welldeveloped, LC well-developed, entire, without pillow at base, twisted with LT, LI absent, LT well-developed, entire, PU present, with hook, PV well-developed, entire, elongate, with simple spines. Photophores: organ of Pesta present, dermal organs absent. Arthrobranchs: posterior lobe on somite XII lamellar, anterior lobe on somite XIII dendritic.
Gardinerosergia gen.n. urn:lsid:zoobank.org:act:887F6724-DB45-4030-AE6A-D4DB0 A684572 Diagnosis: Integument firm, frontal margin of rostrum oblique, no supraorbital or hepatic teeth, hepatic protrusion forming barb. First segment of A I elongate, $1.5 times as long as 3d segment; distoventral end of 3rd segment in males without process; distal tooth of scaphocerite reaching or overreaching end of blade; maxilliped I developed, 3-4-segmented, maxilliped III, 2.0 times as long as Cp, not dimorphic sexually, dactyl not subdivided into specialized subsegments; pereopods I-II ischia without strong movable spines and distally curved tooth; pereopod II with merus lacking protrusion and chela bearing equal fingers, without long setae; chela of pereopod III without strong curved spines and long setae on propodus, fingers subequal; pereopods IV-V present, 6-segmented; pereopod V setose along both margins; uropodal exopod with outer spine, proximal segment not setose along outer margin. Male clasping organ: developed, with 1-7 serrated bristles in an unordered heap, tubercle present. Petasma: LA well-developed, LC well-developed, without pillow at base, twisted with LT, LI well-developed, straight, slender, LT well-developed, entire, PU present, with hook, PV well-developed, entire, elongate, without hooks, suckers, spines, or apical lashes. Photophores: dermal, as opaque spots, small, not fused, position not fixed; arranged in 2 rows on carapace, in 1 row on scaphocerite and uropodal exopod; positioned close to central axis of scaphocerite and close to margin of uropodal exopod. Arthrobranchs: both posterior lobe on somite XII and anterior lobe on somite XIII dendritic.
Neosergestes Judkins, Kensley, 2008 Diagnosis: Integument firm, frontal margin of rostrum oblique, supraorbital tooth absent, hepatic protrusion forming tooth. First segment of A I not elongate, ,1.5 times as long as 3d segment; distoventral end of 3rd segment in males without process; distal tooth of scaphocerite reaching or overreaching end of blade; maxilliped I developed, 3-4-segmented, maxilliped III.2.0 times as long as Cp, not dimorphic sexually, dactyl subdivided into 6 specialized subsegments; pereopods I-II ischia with strong movable spines and distally curved tooth; pereopod II with merus having protrusion and chela bearing equal fingers, without long setae; chela of pereopod III without strong curved spines and without long setae on propodus, fixed finger subequal to dactyl; pereopods IV-V present, 6-segmented; pereopod V setose along one margin; uropodal exopod without outer spine, proximal segment entirely setose along outer margin. Male clasping organ: developed, without serrated bristles, tubercle present. Petasma: LA well-developed, LC well-developed, without pillow at base, twisted with LT, LI well-developed, straight, inflated, LT well-developed, divided, PU present, without hook, PV rudimentary, entire, elongate, without hooks, suckers, spines, or apical lashes. Photophores: organ of Pesta present, dermal organs absent. Arthrobranchs: both posterior lobe on somite XII and anterior lobe on somite XIII dendritic.
Type locality: Eastern Pacific Ocean: Gulf of Panama; Galapagos; and Gulf of California (see details in [41]).
Type species: by present designation, Robustosergia robusta (Smith, 1882) [43]. Etymology: after type species R. robusta (from the Latin 'robusta'' meaning strong, probably an allusion to the exterior which is more robust than in most other sergestids; plus the root 'sergia').
Sergestes H. Milne-Edwards, 1830 Diagnosis: Integument firm, frontal margin of rostrum oblique, supraorbital tooth present, hepatic protrusion forming tooth. First segment of A I not elongate, ,1.5 times as long as 3d segment; distoventral end of 3rd segment in males without process; distal tooth of scaphocerite reaching or overreaching end of blade; maxilliped I developed, 3-4-segmented, maxilliped III,2.0 times as long as Cp, not dimorphic sexually, dactyl subdivided into 6-7 specialized subsegments; pereopods I-II ischia without strong movable spines and distally curved tooth; pereopod II with merus lacking protrusion and chela bearing equal fingers, without long setae; chela of pereopod III without strong curved spines and without long setae on propodus, fixed finger subequal to dactyl; pereopods IV-V present, 6-segmented; pereopod V setose along one margin; uropodal exopod with outer spine, proximal segment not setose along outer margin. Male clasping organ: developed, without serrated bristles, tubercle present. Petasma: LA rudimentary, LC well-developed, divided, without pillow at base, twisted with LT, LI well-developed, straight, slender, LT well-developed, entire, PU present, without hook, PV rudimentary, triangle, without hooks, suckers, spines, or apical lashes. Photophores: organ of Pesta present, dermal organs absent. Arthrobranchs: posterior lobe on somite XII lamellar, anterior lobe on somite XIII dendritic.
Type locality: North Atlantic Ocean near Azores. Species included: Sergestes atlanticus H. MilneEdwards, 1830 [9] Sergia Stimpson, 1860 Diagnosis: Integument membranous, frontal margin of rostrum oblique, no supraorbital or hepatic teeth, hepatic protrusion inconspicuous. First segment of A I elongate, $1.5 times as long as 3d segment; distoventral end of 3rd segment in males without process; distal tooth of scaphocerite not reaching end of blade; maxilliped I developed, 3-4-segmented, maxilliped III,2.0 times as long as Cp, not dimorphic sexually, dactyl not subdivided into specialized subsegments; pereopods I-II ischia without strong movable spines and distally curved tooth; pereopod II with merus lacking protrusion and chela bearing equal fingers, without long setae; chela of pereopod III without strong curved spines and long setae on propodus, fingers subequal; pereopods IV-V present, 6-segmented; pereopod V setose along both margins; uropodal exopod with outer spine, proximal segment not setose along outer margin. Male clasping organ: developed, without serrated bristles, tubercle present. Petasma: LA welldeveloped, LC well-developed, entire, without pillow at base, not twisted with LT, LI well-developed, straight, slender, LT welldeveloped, entire, PU present, with hook, PV well-developed, entire, elongate, without hooks, suckers, spines, or apical lashes. Photophores: dermal photophores and organ of Pesta absent. Arthrobranchs: both posterior lobe on somite XII and anterior lobe on somite XIII dendritic.

Phylogenetic Remarks and New Taxonomy
The phylogeny of the former genera Sergestes and Sergia presented here is the result of simultaneous use of all available morphological characters. There are three groups of characters relating to (1) general decapod morphology, (2) morphology of the male copulatory organs, and (3) morphology of the photophores. Previous attempts to classify sergestid shrimps have mainly focused at one of these character systems resulting in a lack of consensus concerning sergestid systematics. Table 4 shows that only five genera are supported by one type of synapomorphies: either general morphological (Sergia) or sexrelated (Sergestes, Robustosergia, Challengerosergia, Scintillosergia). Most genera are supported by a combination of synapomorphies: general morphological and sex-related (Eusergestes, Cornutosergestes, Neosergestes, Parasergestes, Deosergestes, Allosergestes), general morphological and photophore-related (Gardinerosergia), or sex-and photophore-related (Phorcosergia, Prehensilosergia, Lucensosergia). Use of one type of the characters would apparently not be resulted in a satisfactorily resolved tree. Thus, a simultaneous use of a broad suite of characters including sexand photophore-related characters is necessary for any successful attempt of sergestid phylogeny.
Our analysis does not support the clades Sergia and Sergestes as recognized previously [4], [11]. Instead, 15 separate genera within these former taxa are supported by synapomorphies and Bootstrap analysis. 'Sergia' and 'Sergestes' groups were suggested early, although various experts proposed different taxonomic statuses for them. Initially, the status was generic (first descriptions by Milne Edwards [9] and Stimpson [10]), but later both genera were combined in the single genus Sergestes [4], [11]. Burkenroad [23], [27] suggested a subdivision of the genus into the two subgenera Sergestes s.s. and Sergia, based on differences in types of photophores and pigmentation, which was later defined formally by Yaldwyn [5]. The taxonomic status of both subgenera was raised to generic level by Omori [6]. Omori's classification was subsequently used by most authors but some hinted at the possibility that both genera might have higher taxonomic status, and that the species groups within each genus might deserve the status of valid genera [7] - [8].
Considering the twisted taxonomy of the 'Sergia-Sergestes' group, Vereshchaka [7] - [8] underlined that development of a new classification should be postponed until revisions of the world fauna had been completed and phylogenetic analyses based on a broad set of characters had been undertaken. The situation became more complicated when Judkins, Kensley [18] offered only very brief diagnoses for several new genera. However, the results of the phylogenetic analysis in this work indicate that most of their new genera are valid. In the present work we recognise that Judkins and Kensley's genus 'Sergestes' contains two valid genera (Sergestes Milne-Edwards, 1830 [9] and Cornutosergestes n.gen.) and that the genus Sergia s.s. consists of eight genera (see fig. 8 and Table 2).

Vertical Distribution of Clades
Several evolutionary patterns relating to vertical distribution (benthic, pelagic, etc.) of the genera under consideration can be elucidated based on the consensus tree presented here (Fig. 9). Sicyonella antennata, used as outgroup in the present analysis, is benthopelagic ( Fig. 9, brown). Since other related sergestid taxa (Acetes, Peisos) are also benthopelagic, this optimises as the likely original habitat for sergestids. The genera Lucensosergia and Challengerosergia are characteristic of near-bottom layers above seamounts, continental slopes, and shelves. Most benthopelagic species have local ranges [7] - [8], [45] in the Atlantic, Indian, and Pacific oceans living above shelves, continental slopes, and seamounts. Our phylogeny shows that several lineages have penetrated into the pelagic realm independently from a nearbottom origin. Most clades and genera are typical interzonal migrants living in the low mesopelagic zone (depths 400-800 m) in the daytime and ascending to the epipelagic zone (100-200 m) at night (Fig. 9, light blue). These are species with regional geographic ranges, occupying temperate and/or tropical zones and including the genera Eusergestes, Sergestes, Neosergestes, Parasergestes, Allosergestes, Robustosergia, Phorcosergia, and parts of Gardinerosergia.
Some species of the genera have become epipelagic living in the upper 200 m (Fig. 9, transparent). These are Cornutosergestes and some species of Gardinerosergia. Conversely, Sergia went to the deep bathypelagic zone (800-3000 m deep Fig. 9, blue); this genus shows typical adaptations to deep-sea living such as high fatcontent, membranous integument, small cornea, absence of photophores, etc.

Geographic Distribution
The geographic distribution of the species is mapped on the strict consensus tree in Fig. 10. There is no simple relation between presented phylogeny and the species distribution, so any conclusions concerning 'centres of origin' are very difficult to obtain. Numbers of species occurring in different oceans are similar: 32 in the Atlantic, 37 in the Indian Ocean, and 38 species in the Central and West Pacific. East Pacific is inhabited by a relatively low number of species (14) that may be related to oxygen-depleted conditions recorded in many areas of this region.
The overall pattern of geographic distribution broadly indicates that speciation has occurred mainly in the tropical and subtropical waters of all oceans. Indeed, 68 of 71 recorded species are found in the tropical/subtropical waters of the World Ocean and only the genus Eusergestes inhabits temperate/subpolar waters.
One general pattern concerning distribution is that the species ranges of most species within most genera (except epipelagic Cornutosergestes and bathypelagic Sergia) are restricted to a single ocean (Table 5). These genera are meso-or bathypelagic that indicates importance of sympatric speciation [46] - [47] within these zones. This rule is even more pronounced with respect to benthopelagic species having local species ranges, as is often seen for non-pelagic marine taxa.
One interesting pattern is that most often a single species with panoceanic distribution can be found within each genus (Table 5). However, this does not apply to the epipelagic genus Cornutosergestes, perhaps because the surface anticyclonic gyres in this region result in geographic isolation, thereby preventing wider distribution of shallow-living species. Also, such a pattern does not apply to the bathypelagic species of Sergia, almost all of which are panoceanic probably being distributed by the Great Ocean Conveyor current.
The species with panoceanic distributions are of special interest for molecular studies, since, despite the morphological similarity between populations, cryptic speciation may be involved [48]. If, on the other hand, the panoceanic species are not distinguishable genetically, it would be a challenge to explain their distribution in more detail than we have here.

The Photophores
Most species included in the phylogenetic analysis possess luminescent organs. Species of the genera Eusergestes, Sergestes, Cornutosergestes, Neosergestes, Parasergestes, Allosergestes, and Deosergestes have the organ of Pesta (Fig. 11, in pink), which is a lluminescent modified area of the gastrohepatic glands found within the cephalothorax. The morphology and histology of this organ are different in each of the generic clades (details in [8]) and are therefore of phylogenetic importance.
The genera Prehensilosergia, Scintillosergia, Lucensosergia, Challengerosergia, Gardinerosergia, Robustosergia, and Phorcosergia have dermal photophores (Fig. 11, in green), which are either lens-bearing or lens-less. Lens-bearing photophores may be arranged in two different ways (details in [8]) and are characteristic of the genera Prehensilosergia, Scintillosergia, Lucensosergia, and Challengerosergia. Lens-less photophores (visible in preserved material as opaque spots) are arranged in three different types, each of which is characteristic of Gardinerosergia, Robustosergia, or Phorcosergia.
Sergestid dermal photophores are directed downwards and are important for countershading [49] - [50]. In the bathypelagic zone, countershading is ineffective and all species of the deep-sea genus Sergia have no photophores.

Behavioral Strategies: Offensive Versus Protective
The behaviour of sergestid species is not well known because most of them are mesopelagic and difficult to observe. However, visual observations were available for Lucensosergia lucens [51], Eusergestes similis [52], and the closely related Acetes sibogae [53]. The data indicate that the shrimps capture prey by combined actions of the first three pairs of pereopods and the third maxillipeds before transferring it to the more dorsal second maxillipeds. Simultaneously they move using the pleopods (forward movement) or uropods (escape backwards). The morphology of these appendages reflects the presence of two fundamentally different behavioral strategies: offensive and protective. Some taxa feed on live planktonic animals and have developed a set of structures relating to this feeding mode; this is termed here an 'offensive strategy'. On the other hand, sergestids are themselves preyed upon by larger carnivores like squids and fishes and have developed a set of characters related to protective/ avoidance behavior; this is termed a 'protective strategy'.
Several genera (Neosergestes, Parasergestes, Allosergestes, and Deosergestes) show a set of characters related to the offensive strategy (Fig. 12, red spectrum): (1) they have much enlarged maxillipeds (.2 times as long as first pereopods), which are stretched forward to catch the prey; (2) their uropods may act as rudders (an increased surface area of the blade, which is also enhanced by greater setal coverage), during their slow swimming forward towards prey using the pleopods. Further morphological specializations for feeding on other planktonic animals (offensive strategy) are seen within the genera Neosergestes, Parasergestes, and Allosergestes (Fig. 12, light red). In these species the dactyl of maxilliped III is subdivided into 4-5 very specialized subsegments specialized for catching prey. The genera Neosergestes and Parasergestes (Fig. 12, red) have additional specialized structures for catching prey. The ischia and meri of pereopods I-II in these genera have various teeth, spines, and protrusions that may be used for the retention of prey. Specializations are also present in the genera Allosergestes and Deosergestes (Fig. 12, orange) which seem to be related to an offensive strategy; the chelae of pereopods I-II bear strong teeth and/or elongated setae, which may replace the dactyl functionally.
The remaining clades (the genera Eusergestes, Sergestes, Cornutosergestes, Prehensilosergia, Scintillosergia, Lucensosergia, Challengerosergia, Gardinerosergia, Robustosergia, Phorcosergia, and Sergia) show specializations relating to the protective strategy (escaping predators) (Fig. 12, the green spectrum). They have uropods that are buttressed by a strong tooth and with a reduced number of setae. These uropods are efficient as oars when the shrimp jumps backwards to escape predators. A streamlined body lacking protruding teeth or spines (e.g., hepatic or supraorbital spines) would aid in their escape from predators. Such a morphology is found in the clade which includes Prehensilosergia, Scintillosergia, Gardinerosergia, Robustosergia, Phorcosergia, and Sergia (Fig. 12, green).
Finally, the genera Robustosergia, Phorcosergia, and Sergia (Fig. 12, dark green) show a further advance in the development of a protective strategy. Here, the large size and dense setal coverage of the scaphocerites (lateral blades of antennae) suggest they act as rudders during backward jumps generated by the use of the oartype uropods.
Both the protective and offensive strategies are also related to swarming behavior. All commercially important species (Lucensosergia lucens, Eusergestes similes, Eusergestes arcticus) belong to the ''protective strategy'' group where avoidance of predators is favoured. All species of this group (except for a few rare ones) are numerous in plankton samples, which suggests they aggregation in  Diversity and Phylogeny of the Former Genera Sergestes and Sergia  shoals. Indeed, the maximum concentration observed via underwater camera in aggregations of L. lucens may reach 6 ind/m 3 [54]. The species that favour offensive strategies are regularly recorded but are always few in numbers, which suggests that shoaling behavior is not the case here. Instead, their relative rarity, small size, and remarkable transparency may protect these species from carnivores and thereby explain why no protective morphological structures are present.

Conclusions
There are always subjective aspects of basing a classification on a phylogeny. The most important point is that any suggestion should always be based on a phylogeny using as much data as possible. We tried both to use results of our analysis of three sets of characters and to keep conservative approach as much as we could: wherever possible, we did not erect new taxa and did not confuse sergestid taxonomy any more. All recognized genera are monophyletic, supported by synapomorphies and significant Bootstrap values. Our analysis does not support the genera Sergia and Sergestes sensu Hansen and Omori; instead, it confirms validity of 6 previously established genera [18] and supports 8 new genera. Future work, for example involving molecular data (now very scant in the Gene Bank), will test the validity of suggested phylogeny and thereby the classification. Phylogenetic analysis of the other sergestid genera Acetes, Petalidium, Sicyonella, Peisos may also help in better understanding of the sergestid phylogeny.
Our studies revealed the following phenomena: (1) independent phylogenies of 3 sets of characters (general decapod morphology, male copulatory organs, and photophores); (2) existence of two prominent strategies, protective and offensive, expressed both in morphology and behaviour at generic level; (3) biogeographical rule for mesopelagic genera ''a single panoceanic species -most species with local distribution''. Molecular studies on these panoceanic species may help to understand whether they really represent a single species-level taxon or they are conglomerates of two or more species.
Further studies will show how widely these phenomena are spread among other taxonomic groups.