Spinneret Spigot Morphology in Synaphrid Spiders (Araneae, Synaphridae), with Comments on the Systematics of The Family and Description of A New Species of Synaphris Simon 1894 from Spain

Abstract We describe for the first time the spigot morphology of two synaphrid species (one of each of two synaphrid genera, Synaphris and Cepheia) as well as the morphology of the respiratory system of Synaphris. We also provide a taxonomic description of a new species of Synaphris from Spain, including detailed information about its morphology. This new species is known only from males, and it might belong to the so-called letourneuxi species group. Some morphological features proposed as synapomorphies for the genus Synaphris and/or the Synaphridae are questioned and discussed. Putative synapomorphies proposed here include a distinct constriction on the tarsus-metatarsus joints; a cheliceral keel ending in a strong promarginal cheliceral tooth; scarce number of maxillary setae; distal maxillary setae clavate; and a characteristic palpal morphology, comprising a distinctive tibial morphology, a modified cymbium with two separate areas, a palpal dorsal translucent expansion of the embolar base, a retrolateral paracymbium, a reduced furrow separating the major ampullate field from the piriform field, and the retention of at least one triad spigot in males. Refuted synapomorphies are the metatarsal subdistal anastomosed lyriform organ, the notched tibial trichobothrial base, and the tarsal pseudosegmentation. We also discuss the phylogenetic placement of the family, suggesting a close relationship to the araneoid Cyatholipidae.


INTRODUCTION
The recently erected araneoid spider family Synaphridae Wunderlich 1986(Marusik and Lehtinen, 2003, see also Schü tt, 2003 comprises two genera of minute spiders: the monotypic Cepheia Simon 1894, from southern Europe, and Synaphris Simon 1894, with seven species described from the Canary Islands, Croatia, Ukraine, Turkmenistan, and Egypt ( fig. 1; Platnick, 2006). Two new Synaphris species and a third synaphrid genus from Madagascar are currently being described (Miller, in press). Recent studies (Miller, in press;Marusik et al., 2005;Marusik and Lehtinen, 2003;Schü tt, 2003) have proposed putative synapomorphies in support of the monophyly of the family and/or of the genus Synaphris. The familial synapomorphies include a cheliceral keel that ends in a single strong prolateral tooth (Wunderlich, 1986(Wunderlich, , 1995''tooth-like projection'', Schü tt, 2003; also ''process with pointed tip'', Marusik and Lehtinen, 2003); presence of paracymbium; cymbium modified into a rounded lobe; palpal tibia pressed against the cymbial base and with two pits (Marusik and Lehtinen, 2003: 147, fig. 17; see Discussion below); a flattened and transparent extension of the embolic base called lamella; absence of leg spines; tarsi longer than metatarsi; a notched tibial trichobothrial base; anastomosing lyriform organs on subdistal metatarsi (all features above from Marusik and Lehtinen, 2003); embolar and cymbial conductor; and pseudosegmentation of tarsi (Marusik et al., 2005). Most of the features listed above were observed in only one or two Synaphris species and generalized for the family. No detailed study of the morphology of Cepheia longiseta and of most Synaphris species has ever been done, and their spinneret spigot morphology remains undescribed. Given the importance of spigot morphology for reconstructing the cladistic relationships of araneoid families (e.g., Griswold et al., 1998Griswold et al., , 2005 the study of this character system should contribute to understanding the phylogenetic position of synaphrids. In this paper we describe for the first time the spigot morphology of two synaphrid species (one of each of two synaphrid genera; see also Miller, in press, for the spinneret spigot morphology of three recently described synaphrid species). We also provide a taxonomic description of a new species of Synaphris from Spain, including detailed information about the external morphology and the tracheal system. We discuss some of the morphological features of synaphrids and their potential value as systematic characters. The new species is the first record of the genus Synaphris for the Iberian Peninsula. In addition, we report the first records of Cepheia from Portugal and the Baleares Islands. Madagascar, see Miller (in press). For Synaphris, each location represents one described species. For Cepheia, all locations represent the geographical distribution of Cepheia longiseta.

MATERIAL AND METHODS
Methods of study follow Hormiga (2003). Specimens were studied in 75% ethanol using a Leica MZAPO stereomicroscope. For observation of respiratory structures, the abdomens of two specimens were bisected horizontally and digested with SIGMA Pancreatin LP 1750 enzyme complex in a solution of sodium borate, prepared following the concentrations described by Dingerkus and Uhler (1977) as modified in Alvarez Padilla and Hormiga (in prep.). Some of the tracheal system preparations were stained with an aqueous solution of chlorazol black as described in Hormiga (1994). The bisected abdomen was left in the enzymatic solution overnight and at room temperature. After the enzymatic digestion, the specimens were transferred to distilled water for observation. All measurements are in millimeters. Carapace height was measured at the highest point, from the carapace lateral edge, not from the sternum. Abdominal height, length, and width were measured at the maximum points. To account for length variation, measurements are expressed first as the length of the described specimen, then as the range of all observed specimens (in parentheses, table 1). After dissection, palps were cleared in clove oil. Palp drawings were made with a camera lucida attached to a Leica DMRM compound microscope. For SEM images, the specimens were critical-point dried and sputter-coated with gold-palladium. Images were taken with a LEO 1430VP microscope at the Department of Biological Sciences (GWU) SEM facility. Species descriptions and measurements follow Lopardo (2005) and Marusik et al. (2005). Leg formula refers to the relative length of legs. Two legs are considered equally long when their range of variation highly overlaps, even if their averages are slightly different. We comment on and propose some changes in the system of names used by Marusik et al. (2005). Study specimens will be deposited in the American Museum of Natural History (AMNH, New York), the California Academy of Sciences (CAS, San Francisco), the Museum of Comparative Zoology (MCZ, Harvard University), and the Museo Nacional de Ciencias Naturales de Madrid (MNCN, Spain). For abbreviations used through figures and text see appendix 1.
ETYMOLOGY: The species epithet is an arbitrary combination of letters.

5739-A).
Cepheia longiseta (Simon, 1894) figures 63-80 DESCRIPTION OF SPIGOT MORPHOLOGY: Male: Colulus large, fleshy, triangular, about half length and width of ALS, with three setae (figs. 63, 64). ALS (figs. 65, 66) with one MAP spigot, accompanied by a nubbin and a tartipore, separated by weak (almost nonexistent) furrow from PI field. PI field, on external side of ALS, contains three PI spigots with reduced bases, posterior PI spigot base larger. PMS (figs. 67, 68) with two spigots of similar morphology, chemosensory seta (can be confused with spigot) located anteriorly; externally, its base deepens around shaft. PLS (figs. 69-71) with two spigots of slightly different morphology clumped in same field. Internal one with rounded, larger base and more cylindrical shaft, external one with oval base and tapering shaft. Short, thick chemosensory seta (can be confused with a small spigot) located more basally on internal side of distal PLS segment.

SPINNERET MORPHOLOGY AND WEBS IN SYNAPHRIDAE
Males and females of C. longiseta and S. saphrynis present one major ampullate spigot on their ALS that is separated from the piriform field by a weak furrow, and it is accompanied by a nubbin and a tartipore. In C. longiseta, the PI field presents few piriform spigots, the posterior one with enlarged and more distinctive base. Piriform spigots with different size bases seem to also occur in other spiders (Martín J. Ramírez, personal commun.).
The presence of only one spigot on the PMS of S. saphrynis and two morphologically similar PMS spigots in C. longiseta (not including the cylindrical spigot in females) suggest that these spigots are either minor ampullate silk gland or aciniform silk gland spigots, although these spigots do not necessarily serve the same silk glands in each species. A single spigot is also present in the two recently described Synaphris species and the new synaphrid genus from Madagascar, which has been assigned to a minor ampullate gland (Miller, in press), in which case it would be homologized to that of S. saphrynis as well. Unfortunately, external morphology alone is not sufficient to discriminate between these two spigot types. In the case of C. longiseta, the two PMS spigots most likely correspond to aciniform gland spigots, as the presence of two minor ampullate spigots in araneoid adults is rare. Also, the anterior chemosensory seta in the PMS of C. longiseta has been previously considered a median-anterior minor ampullate silk gland spigot (Schü tt, 2003), although its morphology is similar to the chemosensory seta on the female palp tip (see Lopardo and Hormiga,in press,fig. 12).
The PLS in both S. saphrynis and C. longiseta have two types of spigots in addition to the cylindrical spigot found only in females. If one of these two is an aciniform spigot, at least one spigot from the araneoid triad would have been retained. Consequently, the PLS retained the combination of either one aciniform plus one triad spigot (flagelliform or aggregate), or one aggregate plus flagelliform (a ''triplet'' of two spigots) and no aciniform. As is the case of the singular PMS spigot in Synaphris, external morphology alone is not sufficient to discriminate between these two spigot types.
The weak demarcation of the major ampullate field from the piriform field on the ALS of Synaphris saphrynis and Cepheia longiseta (coded as absent in Cepheia by Schü tt, 2003; see also Miller, in press) contradicts the evidence supporting Synaphridae as a member of Araneoidea. Should the monophyly of Synaphridae and its current placement within Araneoidea hold, then Synaphridae would become the only araneoid family with a weakly demarcated major ampullate field (as proposed by Schü tt, 2003), a condition that is common outside Araneoidea (Griswold et al., 2005). Based on the morphology of PLS spigots, we could describe synaphrid spiders as having retained at least one of the three triad spigots on their PLS in both sexes.
The web of only one species of synaphrids has been reported. Synaphris lehtineni builds a small, thin sheet web underneath stones in hollow depressions where they live (Marusik et al., 2005: 129). The web building behavior of S. lehtineni and the rest of synaphrid species is still unknown. The presence of only one AG and one FL spigot on the PLS has also been reported for the linyphiid Linyphia hortensis (Schü tt, 1995), which spins a small sheet web composed in part by sticky silk (at least for juveniles, Benjamin and Zschokke, 2004), although the details of the sticky thread structure have not been studied. Given that one of the alternative interpretations of the PLS spigot composition in synaphrids is the retention of a triplet consisting of one AG and one FL spigot, then it is possible that these spiders also spin webs that include some kind of sticky silk.  Marusik and Lehtinen (2003: 150) divided Synaphris into two separate species groups, although they did not provide an explicit phylogenetic justification for the monophyly of these two groups. Although the diagnostic characters for these two species groups seem sound, the monophyly of the two groups remains untested, and species membership to the species groups seems to be based on overall similarity. The letourneuxi group includes those species with palpal embolic lamella, two distal apophyses (the ventral one bipartite), a tegular groove accompanying the distal portion of the embolus, coiled copulatory ducts in females, and the absence of an epigynal depression. Species of this group are distributed from the Mediterranean region to Central Asia. The calerensis group includes those species without embolic lamella (although, as these authors suggested, this feature should be carefully checked), two entire distal apophyses, no tegular groove, and an oval depression in the epigynal area. Species of this group are endemic to the Canary Islands. The presence of an embolic lamella, a lower apophysis bipartite, and the tegular furrow in Synaphris saphrynis, n.sp. suggest that this species belongs to the letourneuxi group.

PALPAL MORPHOLOGY OF SYNAPHRIS
The palpal morphology of Synaphris saphrynis, n.sp. is very similar to that of S. letourneuxi (Simon, 1884), the type species; therefore, its generic placement is rather straightforward, although it has not been phylogenetically tested. The male palp of Synaphris is characterized by a cymbium divided into two separate areas: a sclerotized one bearing setae and a glabrous, membranous one, the latter one on the prolateral side. The retrolateral margin of the cymbium in Synaphris presents a notch delimitating a basal expansion, with cuticular differentiation, which most certainly is a paracymbium ( fig. 60; Marusik and Lehtinen, 2003;Miller, in press).
The embolus base is extended into a membranous ''lamella'', and the tegulum is flat and oval. Homology statements regarding palpal sclerites and apophyses are notoriously difficult to test (e.g., Griswold et al., 1998;, sometimes even at the primary stage of the homology proposition. The tegulum of Synaphris saphrynis is distally extended into a large and less sclerotized area, which bears several apophyses and a longitudinal fold that houses the distal region of the embolus (figs. 50, 59). There are no clearly delimited or membranous attached tegular sclerites (such as those often found on the tegulum of many araneoids), except for the embolus and the membrane that arises from the membranous stalk that connects the tegulum to the embolus base (the socalled embolic ''lamella'').
In their detailed description of the male palpal morphology of Synaphris lehtineni (Marusik et al., 2005) and S. orientalis (Marusik and Lehtinen, 2003), the authors use a rich terminology to name and label the various tegular processes (the palpal morphology of these two species is very similar to that of S. saphrynis, n.sp.). Some of the anatomical terms used by Marusik and coworkers have a long history in palpal morphology (e.g., ''terminal apophysis''). Although the choice of names of these morphological features implies homology to at least some of the equally named sclerites in other taxa, these authors do not discuss the potential correspondence of the palpal sclerites of Synaphris with those of other araneoid taxa. Interestingly, Marusik et al. (2005) did not label any of the tegular sclerites as a ''conductor'', but mentioned that the terminal part of the embolus lies in a ''tegular groove'' and that such tegular groove would be unique (autapomorphic) within Araneoidea. These authors also provided names for the tegular processes: ''subterminal apophysis'' (the one at the end of the conductor) and ''terminal apophysis'' (''Cap'', or conductor apophysis). Both names have been used by other araneologists for sclerites in araneoid taxa. We propose what is perhaps a simpler alternative, namely that the distally extended tegular area of Synaphris is a homolog of the araneoid conductor and that the ''subterminal'' and ''terminal'' apophyses of Marusik et al. are processes of the conductor. As is often the case, this conductor partly covers (in a groove or fold in this particular case) the distal region of the embolus. The presence of processes or apophyses on the conductor has been documented in other araneoid taxa (e.g., in Synotaxidae and Anapidae; see for example Platnick and Forster, 1989: fig. 55). The palp of S. saphrynis, n.sp. also exhibits these two separate sclerotized and distinct conductor apophyses.
Metatarsal subdistal anastomosed lyriform organs (Marusik and Lehtinen, 2003: 147, 150). This special lyriform organ has not been clearly defined nor properly documented and illustrated. A similar organ is present in S. saphrynis, n.sp. (figs. 17, 18), as a dorsal band of anastomosed ridges (compare it to the lyriform organ on tibia-metatarsus joint in fig. 26). This organ is also present in other symphytognathoids such as anapids and mysmenids (LL and GH, personal obs.) and seems to also be present in theridiids (e.g., Agnarsson, 2004: fig. 84E; also Jeremy Miller, unpubl. data), theridiosomatids and even in uloborids (Jeremy Miller, unpubl. data). In light of the mentioned evidence this organ can no longer be accurately described as ''a pattern that is not known in other higher Araneoidea'' (Marusik and Lehtinen, 2003: 147). Furthermore, the presence of this organ in Uloboridae suggest that is it widespread at least within Orbiculariae, although might be present in other non-orbicularian spiders as well. In this paper we illustrate this organ for the family, for future reference. Nevertheless, a putative synapomorphy for the family seems to be the constriction on the tarsi-metatarsi joint on every leg and both sexes, which we do not know for any other Araneoidea, except for the pimoid genus Nanoa Hormiga, Buckle and Scharff 2005 (Hormiga et al., 2005: 252;GH, personal obs.).
The presence of a cheliceral keel that ends in a strong promarginal cheliceral tooth seems to also define the family (Wunderlich, 1986(Wunderlich, , 1995Schü tt, 2003), although it is not unique to Synaphridae. The concept of ''cheliceral keel'' has been previously used to name different structures. The following discussion stems from the premise that anatomical terminology should reflect hypotheses of homology. This cheliceral structure has been correctly defined as ''the inner side of the paturon showing a distinct edge over the entire length'' by Schü tt (2003: 141), but it has also been defined as a ''serrated ridge, truncated outgrowth of the cheliceral distal margin'' (Ubick et al., 2005: 266, fig. 72.41; our italics). Alternatively, the cheliceral keel (as defined by Schü tt, 2003) has been referred as a ''process with pointed tip'' by Marusik and Lehtinen (2003), but also as a cheliceral ''lamina'' for the Tetrablemmidae by Lehtinen (1981). Furthermore, the term ''lamina'' has been equated to the cheliceral ''lamella'', defined as a ''sclerotized ridge on the cheliceral margin or mesal surface'' (i.e., Schü tt's keel; Ubick et al., 2005: 266, fig. 72.40). So far, the terms ''keel' ' (sensu Schü tt, 2003), ''lamina'', and ''cheliceral lamella'' would appear to be comparable, except that the ''cheliceral lamella'' has been previously defined as a different structure. Kaston (1978: 269) defined the cheliceral lamella as ''a triangular plate on the promargin of the cheliceral fang furrow in some spiders, resembling a broad tooth, and forms a kind of chela with the fang''. Moreover, the term (cheliceral) ''lamella'' sensu Kaston (1978) has been widely used in descriptions of haplogyne spiders, although under the name of ''lamina'' (e.g., Platnick et al., 1991; in the Gnaphosidae, Platnick, 1975;Medan, 2001; and in the Zodariidae, Szüts and Jocqué, 2001). Because the confusing term cheliceral ''lamina'' has been used to refer to either the cheliceral ''lamella'' or the cheliceral ''keel'', we think that, given the misunderstanding, the term lamina should be avoided whenever possible, and replaced by its respective synonyms. The cheliceral ''keel'' should be defined as a distinct sclerotized ridge on the inner side of the paturon over its entire length. The term cheliceral ''lamella'' has been clearly defined by Kaston (1978). Regarding the ''other keel'' (the ''serrated ridge, truncated outgrowth of the cheliceral distal margin'' [Ubick et al., 2005]), the term ''serrated distal ridge'' seems appropriate.
Potential synapomorphies (related with the mouthparts) for at least Synaphris could be the scarce number of maxillary setae around the labrum and labium and the clavate setae on the distal maxillae ( fig. 5). These features seem to be rare among spiders (Martín J. Ramírez, personal commun.) The dorsal translucent expansion of the embolar base (the embolic lamella) is a putative synapomorphy for at least Synaphris, as no embolic lamella seems to be present in the palp of Cepheia longiseta (Thaler and Noflatscher, 1990 figs. 25-27;Lopardo and Hormiga, in press; contra Marusik and Lehtinen, 2003). This embolic lamella is also absent in males of mysmenid and anapid spiders (LL, unpubl. data). It seems that other Synaphris species, besides S. orientalis, S. lehtineni, and S. saphrynis, n.sp., might have an embolic lamella (e.g., S. letourneuxi and S. dalmatensis: i.e., the letourneuxi group, Marusik and Lehtinen, 2003; also the two new Malagasy Synaphris species, Miller, in press). SEM studies are needed for the calerensis group (S. calerensis, S. agaetensis and S. franzi). The homology of the embolic lamella with the theridiid conductor (at least in Theridion, arising from the embolar base, Marusik and Lehtineni, 2003) remains untested. Also, the term ''lamella'' is widely used for structures that are unlikely to be homologous to the membrane found in Synaphridae (e.g., in Linyphiidae the lamella is a sclerite attached to the radix: the ''lamella characteristica''; see Hormiga, 1994). We prefer to call the embolic ''lamella'' a ''dorsal translucent expansion of the embolar base'' until a proper homology statement is proposed.
The shape of the male palpal tibia, which is rounded retrolaterally and pressed against the cymbial base, appears to be synapomorphic for the family. However, the presence of two pits on the palpal tibia (Marusik and Lehtinen, 2003: 147, fig. 17) has not been properly described, but they appear to be two small depressions in the distal surface of the tibia that contain the cymbium. We cannot find such depressions in S. saphrynis, n.sp. Conversely, these cavities on the tibia seem to be absent in Cepheia longiseta (LL, personal obs.). For the discussion of the embolar and cymbial conductors, see Palpal Morphology of Synaphris.
The notched tibial trichobothrial base does not seem to be a synapomorphy for the family (as proposed by Marusik and Lehtinen, 2003, based on the study of S. orientalis; see also Miller, in press). A notched trichobothrial base is absent from the tibiae and the metatarsi of S. saphrynis, n.sp. Instead, we observed a smooth crescent-shaped, hooded trichobothrial base in both tibiae and metatarsi, as it is known from some other Araneoidea (see Miller, in press). The basal hood presents one lateral ridge on each side, which seems synapomorphic for the family as well. The same smooth trichobothrial base seems to be present in S. lehtineni and the Malagasy Synaphris species, further suggesting the autapomorphic nature of such condition in S. orientalis.
Another family-level diagnostic character is the pseudosegmentation of the tarsi (Marusik et al., 2005). None of the tarsi that we have studied using SEM for S. saphrynis, n.sp. is pseudosegmented. Miller (in press) found tarsus pseudosegmentation in the three studied synaphrid species. This character may define clades within the family.
Tarsi and metatarsi relative lengths seem to vary across symphytognathoids. According to Marusik and Lehtinen (2003) ''the tarsi of all Synaphridae are longer than the metatarsi.'' A closer examination of the relative length of these leg articles across synaphrid taxa reveals that it is not always the case. Furthermore, the statistical significance of ''longer'' versus ''shorter'' is neither calculated nor discussed. In their description of S. orientalis, only measurements of the lengths of segments on leg I are provided, and in this species tarsus I does seem to be longer (ta 0.29/mt 0.24), although no other measurements of additional legs are given. For tarsus/metatarsus I, tarsi also seem to be longer in S. calerensis (0.22/ 0.19), S. franzi (0.23/0.18), and the females of S. dalmatensis (0.30/0.27; Wunderlich, 1980Wunderlich, , 1987. Metatarsi seem to be longer than tarsi in S. agaetensis (leg I: ta 0.18/mt 0.23), the male of S. dalmatensis (leg I: 0.21/0.24; Wunderlich, 1980, 1987), andS. letourneuxi (leg I: 0.29/0.32;II: 0.29/0.32;III: 0.24/0.29;IV: 0.27/0.29;Brignoli, 1970). In the case of S. lehtineni, according to table 1 in Marusik et al. (2005), the tarsi seem to be slightly longer than metatarsi only for leg I, whereas tarsi of the other legs seem equal in length to the corresponding metatarsi. In S. saphrynis, n.sp. the tarsi seem to be slightly longer than metatarsi on legs I and II, but equally long on legs III and IV (table 1). In the case of Cepheia longiseta, in both sexes the tarsi and metatarsi are of equal length, except for the leg III, where the metatarsus is shorter (0.20/0.16) (Brignoli, 1970). In conclusion, as no statistical tests were carried out in any of these cases, we suspect that if such differences in length were tested, they would turn out to be not significant. As Simon (1894: 589) stated in his original description of Synaphris, the tarsi and metatarsi are almost equally long.
Regarding the spinneret spigot morphology, the reduction of the furrow separating the ampullate and piriform fields seem to be synapomorphic for Synaphridae. Also, while not unique to Synaphridae, the potential retention of triplet spigots in adult synaphrid males might be characteristic for the family (see Miller, in press). Finally, if the retention of two of the three triad spigots hold, the loss of one aggregate spigot in the PLS triplet might be synapomorphic for the family as well, as proposed by Miller (in press).
In conclusion, putative synapomorphies for Synaphris and/or Synaphridae include: a distinct constriction on the tarsus-metatarsus joints; a cheliceral keel ending in a strong promarginal cheliceral tooth; scarce number of maxillary setae; distal maxillary setae clavate; and a characteristic palpal morphology, comprising a distinctive tibial morphology (rounded retrolaterally and pressed against the cymbial base), a modified cymbium with two distinct areas (a sclerotized one bearing setae, and a glabrous membranous one), a palpal dorsal translucent expansion of the embolar base (embolic ''lamella''), a retrolateral paracymbium, and a reduced furrow separating the MAP from the piriform field. Refuted synapomorphies for the genus or the family are the metatarsal subdistal anastomosed lyriform organ, the notched tibial trichobothrial base, and the tarsal pseudosegmentation. Marusik and Lehtinen (2003) provided a detailed discussion about the potential placement of Synaphridae within Araneoidea. They concluded that synaphrids were not nested within mysmenids and discussed the exclusion of the Synaphridae from the Symphytognathoidea (contra Wunderlich, 1986, 1995and Schütt, 2003. They also concluded that the family is not related to micropholcommatids, and although the shape of the lateral surface of the palpal tibia (developed into a rounded lobe) is present in most theridiids, ''…this group [Synaphridae] cannot be placed in any family to which it has been assigned during its history (Theridiidae, Mysmenidae, Symphytognathidae, and Anapidae), because it lacks the generally accepted synapomorphies of all these groups'' (Marusik and Lehtinen, 2003: 147). They also mentioned a similarity with the Cyatholipidae based on the position of the tarsal organ (ca. 0.36), but they did not propose a formal hypothesis of relationship.

THE POSITION OF SYNAPHRIDAE
Surprisingly Marusik and Lehtinen (2003: 151) proposed that the genus Synaphris is related to Theridion (Theridiidae): ''It seems that Synaphris is related to or at least more similar to Theridiidae (Theridiinae, Theridion). Common (similar) characters are as follows: 1) a circular base of the embolus, 2) tip of the embolus terminal in relation to the tegulum, 3) internal 'paracymbium', 4) circular margin of the palpal tibia covering basal part of the cymbium and the bulb, 5) the conductor (5 embolic lamella in Synaphris) originating from under the embolus base''. We consider that the sharing of these features might not be unique to Synaphris-Theridion, and therefore do not allow for the suggestion of a close relationship. Furthermore, such a proposal would require one to empirically demonstrate that Synaphris shares the synapomorphies of Theridiidae (see Discussion above). The typical theridiid tarsal comb on leg IV is absent in Synaphris saphrynis.
The position of Synaphridae within Araneomorphae is controversial due to sets of characters that support contradicting placements. On the one hand, the presence of a paracymbium in Synaphris suggests a potential placement within Araneoidea, although not close to Theridiidae, which lacks a basal paracymbium. The anterior respiratory system transformed into tracheae, although present in Caponiidae, within Araneoidea seems exclusive to symphytognathoids (Schü tt, 2003;Griswold et al., 1998) and further supports (or at least does not refute) a placement of the family within Araneoidea. The same conclusion can be reached from the anastomosed lyriform organs in the distal metatarsi, which appear to be present in symphytognathoids as well as in theridiids and uloborids, and might be common within orbicularians. On the other hand, the weak furrow between the major ampullate and the piriform spinning fields (considered absent by Schü tt, 2003) contradicts the placement of Synaphridae within Araneoidea (or suggest instances of homoplasy) and therefore the spinneret morphology might place Synaphridae in a different position (as a nonaraneoid) than previously thought.
The broad posterior tracheal spiracle (i.e., two distant spiracles connected by a furrow), the retention of PLS triplet spigots in the adults males, and the absence of female palpal claw (at least in Cepheia females, LL personal obs.; also in the Malagasy species, Miller, in press) are some of the proposed synapomorphies of the araneoid family Cyatholipidae (Griswold, 2001). The truncated posterior sternum margin and the absence of median apophysis on male palps, while common among symphytognathoids, are also characteristic not only for the cyatholipids but also for its sister group, the family Synotaxidae (Griswold, 2001). Furthermore, a posterior respiratory system similar to that of Synaphris (two main tracheal bundles from two connected atria, with the tracheoles directed mainly anteriorly and reaching into the prosoma, i.e., desmitracheate) has been previously described for Cyatholipidae (Davies, 1978;Forster, 1980Forster, , 1988Griswold, 2001) and for some groups of Linyphiidae (Blest, 1979;Millidge, 1986Millidge, , 1988Hormiga, 1994Hormiga, , 2000. The cyatholipid and linyphiid (mainly, but not exclusively, in the subfamily Erigoninae) tracheal systems include lateral tracheae arising from a less deep atrium (e.g., Forster, 1988: figs. 19-20;Hormiga, 2000: fig. 31G), which seems to be different from that in Synaphris. The apodemal lobes present in cyatholipids (Forster, 1988 figs. 19-20;Griswold, 2001 figs. 12A-C) are also absent in Synaphris. The presence of several of the proposed cyatholipid synapomorphies in synaphrids may suggest a close relationship to this araneoid family. Of course, this conjecture needs to be tested within the context of a cladistic analysis with the appropriate taxonomic sample. The study of new synaphrid species as well as potential outgroup taxa will test the hypotheses of homology that we have described as well as the cladistic position of the family. Wunderlich, J. 1995