EUROPEAN 2003 (LOGUNOV D.V. & PENNEY D. eds.), pp. 349–366. © ARTHROPODA SELECTA (Special Issue No.1, 2004). ISSN 0136-006X (Proceedings of the 21st European Colloquium of Arachnology, St.-Petersburg, 4–9 August 2003)

The : some problems of origin and evolution

Õåëèöåðû ïàóêîâ: ïðîáëåìû ïðîèñõîæäåíèÿ è ýâîëþöèè

S.L. ZONSTEIN Ñ.Ë. ÇÎÍØÒÅÉÍ

Department of Zoology, The George S. Wise Faculty of Life Sciences, Tel-Aviv University, 69978 Tel-Aviv, Israel. email: [email protected] Îòäåë çîîëîãèè, Ôàêóëüòåò åñòåñòâåííûõ íàóê èì. Äæîðæ Ñ. Âèçå, Òåëü-Àâèâñêèé óíèâåðñèòåò, Òåëü- Àâèâ, Èçðàèëü. email: [email protected]

ABSTRACT. Some problems of the origin and evolution of the chelicerae in are discussed. The enlargement and bending of the cheliceral segments combined with the displacement of the fang insertion point ventrad is recognized as the principal cheliceral modification observed in the Araneae. The chelicerae of the and the are regarded as plesiomor- phic in their position and axial orientation, but as apomorphic by their configuration; the latter seems to be an evolutionary compromise caused by burrowing activity. The possibility of an origin of the labidognathous cheliceral construction from the specialized orthognathous chelicerae of liphistiids and mygalomorphs is regarded as improbable.

ÐÅÇÞÌÅ.  ñòàòüå îáñóæäàþòñÿ âîïðîñû ïðîèñõîæäåíèÿ è ýâîëþöèè õåëèöåð ó ïàóêîâ. Îñíîâíûìè ìîäèôèêàöèÿìè õåëèöåð, íàáëþäàåìûìè ó îðòîãíàòíûõ ïàóêîâ, ïðèçíàþòñÿ óêðóïíåíèå è âûãèáàíèå îñíîâíûõ ÷ëåíèêîâ õåëèöåð è ñìåùåíèå òî÷êè ïðèêðåïëåíèÿ êîãîòêà íà èõ âåíòðàëüíóþ ñòîðîíó. Õåëèöåðû Liphistiidae è Mygalomorphae ïðèíèìàþòñÿ ïëåçèîìîðôíûìè ïî çàíèìàåìîìó ïîëîæåíèþ è îñåâîé îðèåíòàöèè, íî àïîìîðôíûìè ïî ñâîåé êîíôèãóðàöèè; ïîñëåäíÿÿ ïðåäñòàâëÿåòñÿ ðåçóëüòàòîì ýâîëþöèîííîãî êîìïðîìèññà, âûçâàííîãî ïåðåõîäîì ê íîðíîìó îáðàçó æèçíè. Âîçìîæíîñòü ïðîèñõîæäåíèÿ õåëèöåð ëàáèäîãíàòíîãî òèïà èç ñïåöèàëèçèðîâàííûõ îðòîãíàòíûõ õåëèöåð ëèôèñòèèä è ìèãàëîìîð- ôîâ ïðèçíàåòñÿ ìàëîâåðîÿòíîé.

KEY WORDS: Araneae, spiders, chelicerae, evolution. ÊËÞ×ÅÂÛÅ ÑËÎÂÀ: Araneae, ïàóêè, õåëèöåðû, ýâîëþöèÿ.

Introduction considered (explicitly or by implication) the orthognathous chelicerae in spiders as precur- Until recently the traditional standpoint treat- sors of the labidognathous type. Starobogatov ing the cheliceral construction in the orthogna- [1985] also used a third category, plagiaxiality, thous spiders as an initial state of the labidogna- a somewhat intermediate cheliceral orientation thous chelicerae was dominant and used in a occurring in the Hypochiloidea. Otherwise, his great number of arachnological publications reconstruction was given in the traditional style: [see Platnick & Gertsch, 1976]. All those works orthognathy–plagiaxial state–labidognathy. 350 EUROPEAN ARACHNOLOGY 2003

Kraus & Kraus [1993] first arrived at the (North American mygalomorphs including repre- conclusion that not only labidognathous cheli- sentatives of the , Microhexura, etc.). cerae, but also the orthognathous cheliceral con- Also examined were some taxa of the Uropygi and struction represented the apomorphic modifica- the . The majority of whip spider tions of a primary type defined as the plagiaxial specimens used, a few females representing two of , and a specimen of Hypochilus variant. In their opinion, this model can be repre- coylei (all from the collection of the American Mu- sented in all main lineages of the order Araneae. seum of Natural History) were kindly lent to me for The cheliceral configurations known for the Li- study through the courtesy of Dr. Norman Platnick. phistiidae, for a number of mygalomorph taxa Most of the figures were prepared specially for possessing shorter chelicerae, and for the repre- this study. Additionally, I had to use some drawings sentatives of the family Hypochilidae, were con- taken from other sources, all these were scanned so sidered as the most similar to the mentioned type. that the output was changed minimally from their Dunlop [1997] supported this hypothesis original form. In a few extraordinary cases some and added a new state, palaeognathy, found in figures were redrawn from the original source. All members of the fossil Paleozoic order Trigono- credits and relevant references are provided. With minor exceptions, the terminology follows tarbida. He suggested that both orthognathy Kaestner [1956], Starobogatov [1985], Shultz and labidognathy can be derived from palaeog- [1990], Kraus & Kraus [1993] and Dunlop [1994, nathy through the plagiaxial type by simple 1996c, 1997]. The subbasal cheliceral segment cor- torsions of basal cheliceral segments, and some responds to the cheliceral coxa of lower , particular states can be explained by reversals. absent in spiders and their Recent relatives [see However, the palaeognath–plagiaxial ver- Dunlop, 1997]. sion appears to be only one of a few possible Some terms as well as the states denoted should ways of explaining the situation observed. A be specified. The term plagiaxial is used only in new attempt to restore the idea to separate the cases when the cheliceral fangs meet each other at approximately 90°, axial orientation of the basal orthognaths and labidognaths from each other cheliceral segments occupies an intermediate posi- was made by Eskov & Zonstein [1990a]. A tion between the forward-directed and downward- little later, I demonstrated another scheme show- directed types; these axes are divergent laterally. ing the independent origin of orthognathous The terms palaeognathy, plesiognathy and neog- and labidognathous chelicerae in spiders at the nathy as well as the derived forms (e.g., plesiogna- Third Eurasian Arachnological Conference. This thous) are used to designate the orthognathous states report was briefly reviewed by Mikhailov found in the , in the Pedipalpi (that [1992], but it has never been published. Despite includes the orders , Uropygi and Schi- the general formal resemblance of both hypoth- zomida) and in the Araneae, respectively. The first term (palaeognathy) was first used by Dunlop [1997], eses, they were based on quite different as- whereas the latter two are newly applied to the sumptions and a number of the significant dif- arachnids. The swollen neognathous chelicerae are ferences between them could be listed. The characterized by the claw insertion point displaced expanded and reworked variant of the 1992 ventrad [Dunlop, 1994: fig. 1]. The least modified version is given here. plesiognathous construction retains many ancestral characters including a trapezoidal basal cheliceral Material and methods segment provided with a toothed vestige of the former immovable finger [Dunlop, 1994: figs 2, 3]. Finally, A noticeable part of the material used for this the palaeognathous cheliceral configuration includes study was lent from the spider collection of the both the fangs more or less displaced, and the gener- Zoological Institute of the Russian Academy of al cheliceral orientation changed from forward- to Sciences, St. Petersburg, in the early 1990s. It in- downward-directed [Dunlop, 1994: fig. 4]. cluded numerous mygalomorph taxa belonging to The term phrynid type applies only to the gener- the families Atypidae, , Actinopodidae, alized subchelate construction occurring in the Re- , Cyrtaucheniidae, , Diplu- cent taxa represented by the Amblypygi, the Uropy- ridae, , Barychelidae and Theraphosidae. gi, the , and found to be applicable also In addition, material was kindly sent to me by Dr. for some members of the fossil order Trigonotarbida Kirill Eskov (Heptathele) and Dr. Frederick Coyle (for details see the corresponding parts of this study). S.L. Zonstein. Origin and evolution of spider chelicerae 351

This term reflects a certain evolutionary level only; Pedipalpi it does not convey any taxonomic context. Plesiomorphically, ancestors of the Tetra- pulmonata presumably possessed three-seg- Results and discussion mented, chelate chelicerae with the movable finger (the future tetrapulmonate cheliceral fang) Origin located retrolaterally; it is the cheliceral con- struction that is shared by the less advanced arachnids, including representatives of the or- The allied groups der Palpigradi adjacent to the tetrapulmonates Most phylogenetic studies agree in placing [Dunlop, 1997; Selden & Dunlop, 1998]. Araneae into the same group with Amblypygi, The orthognathous chelicerae occurring in Uropygi and Schizomida — other or- the representatives of those orders were consid- ders with members also possessing two pairs of ered to be the most archaic cheliceral construc- book-lungs and the clasp-knife-shaped cheli- tion occurring within this pair of taxa. The main cerae1. Together they constitute the Tetrapul- constructive features of this type are as follows: monata Shultz, 1990. Some disagreement oc- (1) contrary to spiders, the chelicerae here are curs only in drawing spiders closer to one or relatively small (cf. Figs 1–8); (2) basal cheli- another subtaxon within this group. ceral segments are trapezoidal and flattened; Until recently Amblypygi were found to be (3) the fang is relatively short and joined to the the closest relatives of spiders [see Platnick & basal segment close to its foremost dorso-api- Gertsch, 1976]. Most often Uropygi s.lat. (The- cal vertex; (4) the toothed part of cheliceral lyphonida + Schizomida) has been considered a furrow representing a vestige of the former sister group of this pair [Grasshoff, 1978; Wey- immovable cheliceral finger is also rather short, goldt & Paulus, 1979; Wheeler & Hayashi, it does not extend to the basal cheliceral mar- 1998]. On the other hand, Shultz [1990] fol- gin; (5) unlike other tetrapulmonate orders the lowed Pocock’s [1893] scheme proposing a fang is more or less toothed — a symplesiomor- closer relationship between Amblypygi and phy shared with more archaic arachnids. Uropygi s.lat.; Araneae was recognized to be an Theoretically, a variant just like this would adelphotaxon of this group. Palaeontological underlie the construction of the neognathous studies also supported the latter model [except spider chelicerae because it appears to repre- Selden, 1996a] and included in the Tetrapul- sent a certain intermediate state between the monata the fossil order Trigonotarbida [Shear chelate construction of more archaic arachnids and the piercing spider type. This intergrade et al., 1987; Selden et al., 1991; Selden & can be derived from the archetype by rotation of Dunlop, 1998]. the basal cheliceral segments around their axes Among recent studies considering or touch- so that the movable cheliceral finger changes its ing upon the problem, some grouped Amblypy- position from anteroretrolateral to anterodor- gi with Uropygi [Giribet et al., 2002], some sal. In some advanced arachnid orders other treat both above hypotheses as equally compet- than higher tetrapulmonates a contrary torsion itive [see Miller, 2003], whereas others do not is present; thus, Pseudoscorpiones and Solif- clearly prefer either of them [Harvey, 2002, ugae share the movable finger displaced com- 2003]. In this study the possibility of drawing pletely or partially ventrad. spiders together with both alternative groups is However, according to Kraus & Kraus also taken into consideration, although the po- [1993], some data allow the assumption that sition of Shultz [1990, 1999] and subsequent orthognathy in the Pedipalpi and the Araneae authors is found to be more substantiated. could have a different origin. They noted that unlike the spiders at least the amblypygids pos- sess a proximal cheliceral apodema arising ret- 1 Within the Arachnida only ricinuleids also possess two-segmented subchelate chelicerae of similar shape rodorsally; correspondingly, there are certain [see Dunlop, 1996c: fig. 11]. differences in the retracting musculature ar- 352 EUROPEAN ARACHNOLOGY 2003

Figs 1–8. Chelicerae — dorsal view and comparative dimensions in some whip spider and spider taxa: 1 — Charinus sp.; 2 — Paraphrynus mexicanus (Bilimek); 3 — Liphistius panching Platnick et Sedgwick; 4 — sp.; 5 — Phyxioschema sp.; 6 — Anemesia karatauvi (Andreeva); 7 — Cteniza sp.; 8 — Hypochilus coylei Platnick. Ðèñ. 1–8. Õåëèöåðû — îáùèé âèä äîðçàëüíî è ñðàâíèòåëüíûå ðàçìåðû â íåêîòîðûõ òàêñîíàõ ôðèíîâ è ïàóêîâ: 1 — Charinus sp.; 2 — Paraphrynus mexicanus (Bilimek); 3 — Liphistius panching Platnick et Sedgwick; 4 — Atypus sp.; 5 — Phyxioschema sp.; 6 — Anemesia karatauvi (Andreeva); 7 — Cteniza sp.; 8 — Hypochilus coylei Platnick. S.L. Zonstein. Origin and evolution of spider chelicerae 353

Figs 9–12. Chelicerae of whip spiders and schizomids: 9 — Paraphrynus mexicanus (Bilimek); 10 — Phrynus marginemaculatus C.L. Koch; 11 — Charinus sp.; 12 — Stenochrus portoricensis Chamberlin [after Tourinho & Kury, 1999: fig. 5]; 9, 12 — retrolateral view; 10, 11 — prolateral view. Ðèñ. 9–12. Õåëèöåðû ôðèíîâ è ñõèçîìèä: 9 — Paraphrynus mexicanus (Bilimek); 10 — Phrynus marginemac- ulatus C.L. Koch; 11 — Charinus sp.; 12 — Stenochrus portoricensis Chamberlin [ïî Tourinho & Kury, 1999: fig. 5]; 9, 12 — ðåòðîëàòåðàëüíî; 10, 11 — ïðîëàòåðàëüíî. rangement. A similar structure is evident in the Trigonotarbida schizomid cheliceral construction (Fig. 23). Trigonotarbids, fossil Paleozoic arachnids, These observations could potentially indicate constitute one more order known for their clasp- the noticeable structural differences between knife-like chelicerae [Shear et al., 1987; Dun- two types of orthognathous chelicerae. lop, 1994]. The order was found to be a sister On the other hand, it should be noted that the group of other tetrapulmonates [Dunlop, 1996b; representatives of many spider families possess Selden & Dunlop, 1998]; at the same time it a similar cuticular cheliceral process, though in shows some relationship to the [Dun- the less developed or in the rudimentary form, lop, 1996c] 2. But despite the proposed posi- and located medially or prodorsally. In the spi- tion, the Trigonotarbida has never been re- der taxa examined it was absent only from the 2 Both Trigonotarbida and Ricinulei share some pos- representatives of higher araneomorphs (Entel- sible synapomorphies as well as the partially or entirely egynae). Both its reduction and overgrown de- hidden chelicerae [Dunlop, 1996c]. The problem is that velopment (as in the Filistatidae) appear to be the last order is generally placed into the same branch with innovative. Therefore, it seems that the pres- the Acari, far away from the Araneae and its relatives [Weygoldt & Paulus, 1979; Shultz, 1990, 1999; etc.]. ence of a large retrodorsal cheliceral process in Moreover, exactly this union is the most concordant clade the Pedipalpi does not belong to the principal shared by a number of the otherwise conflicting arachnid definitive characters; it could be considered an cladograms proposed by different authors [see Wheeler & Hayashi, 1998; Harvey, 2002]. Otherwise, Giribet et al. additional cheliceral modification shared by [2002] consider the union of Trigonotarbida and Ricinulei members of this group. to be more substantiated. 354 EUROPEAN ARACHNOLOGY 2003

Figs 13–17. Chelicerae of trigonotarbids and ancestral pre-mesothele spiders: 13–15 — some modified variants occurring in trigonotarbids [Dunlop, 1997: fig. 4]; 16 — Gelasinotarbus reticulatus Shear, Selden et Rolfe, prolateral view [Shear et al., 1987: fig. 68]; 17 — Attercopus, the same position [Selden et al., 1991: fig. 4c]. Abbreviations: bs = basal cheliceral sclerite; fg = cheliceral fang; sc1 = vestigial first (subbasal) cheliceral sclerite. Ðèñ. 13–17. Õåëèöåðû òðèãîíîòàðáèä è ïðåä-ëèôèñòèîìîðôíûõ ïàóêîâ: 13–15 — íåêîòîðûå ìîäèôèêàöèè õåëèöåð ó òðèãîíîòàðáèä [Dunlop, 1997: fig. 4]; 16 — Gelasinotarbus reticulatus Shear, Selden et Rolfe, õåëèöåðà ïðîëàòåðàëüíî [Shear et al., 1987: fig. 68]; 17 — Attercopus, òî æå [Selden et al., 1991: fig. 4c]. Ñîêðàùåíèÿ: bs = áàçàëüíûé ñêëåðèò õåëèöåð; fg = êîãîòîê õåëèöåðû; sc1 = ðóäèìåíò ïåðâîãî (ñóááàçàëüíîãî) õåëèöåðàëüíîãî ñêëåðèòà. ferred to the same lineage as the Araneae [see paraxial but directed downwards, as in higher Selden et al., 1991; Dunlop, 1996c]. araneomorph spiders (cf. Figs 13–15, 22–25). Like the representatives of Amblypygi and Dunlop [1997] described this as palaeognathy Uropygi, trigonotarbids possessed rather small and supposed that it represents a precursor of orthognathous chelicerae, which were strongly the plagiaxiality in spiders developed later, ac- flattened bilaterally [Shear et al., 1987; Dun- cording to the O. & M. Kraus’ hypothesis, into lop, 1996a]. A few of them also possessed a the orthognathy of most mygalomorphs and the vestige of the third (subbasal) cheliceral seg- labidognathy of the more advanced araneomor- ment that is absent in the extant tetrapulmonate phs, respectively. In this scheme the opposing orders [Dunlop, 1997]. In some trigonotarbid phrynid type represents a synapomorphy shared taxa the general shape of the chelicerae resem- by the Amblypygi and the Uropygi s.lat. In bled the configuration of the mygalomorphs, most respects however, the last construction however, the trigonotarbid chelicerae were appears to be less modified than both men- S.L. Zonstein. Origin and evolution of spider chelicerae 355 tioned earlier (see above). Hence, either trigo- tally, the analogous cheliceral modifications could notarbid and spider chelicerae should be treat- not be caused inevitably by the same evolution- ed as developed independently from each other, ary factors as in the trigonotarbids. The presence or the phrynid type shared by other tetrapulmo- within the Trigonotarbida of some less special- nate arachnids has to be considered a reversal. ized forms with chelicerae weakly distinguish- The last option seems to be less preferable than able from the phrynid type3 (Fig. 16), provides the first for the reasons given below. reliable evidence in favour of this viewpoint. First, such a reversal should affect not one but several characters at once: the general con- Pre-mesothele fossil spiders figuration of the basal cheliceral joint, the shape Almost all Paleozoic ‘orthognath’ and ‘labi- of the cheliceral furrow, the fang shape and dognath’ spiders described prior to the end of the configuration, the locus of the fang insertion 1980s, later were shown to either belong to other point, as well as a general axial orientation of arachnid orders or to be outside the chelicerates the chelicerae. Second, the mentioned phrynid altogether [Starobogatov, 1985; Eskov & Zon- state corresponds more to the transitional vari- stein, 1990a,b; Selden et al., 1991, 1999]. Unlike ant joining the palpigrade and spider condi- those taxa, the Attercopus was de- tions, as was noted above. The last event could scribed possessing both with func- be explained more parsimoniously by assum- tional spigots and the cheliceral glands, which ing this intermediate structure as the most confirmed its membership in the Araneae [Selden archaic construction, common for tetrapulmo- et al., 1991]. These authors specially noted for nate arachnids. Attercopus the first appearance of a naked cheli- If the union of the Trigonotarbida and the ceral fang representing a typical spider feature Ricinulei actually represents a clade, not a grade, (in all other tetrapulmonates it is plesiomorphi- one more contradiction is evident. Although the cally covered with setae or hairs, cf. Figs 9–16). ricinuleids also possess the downwards-hang- Its uniqueness in certain characters (the absence ing chelicerae, their construction seems to be of the trichobothria, etc.), makes this taxon distinctly more archaic than that of the trigono- distinct from all other spider groups [Selden et tarbids. Besides the mobile cheliceral finger, it al., 1991: fig. 3; Selden, 1996a: fig. 5a]. possesses also the immobile one, though in a With the exception that the chelicerae of reduced form. The last detail has been lost in Attercopus were noticeably shorter, they re- the trigonotarbid chelicerae which calls for one semble in general those of the liphistiomorphs: more reversal in the ricinuleids. Therefore, po- the basal segment with an arched upper part and tentially this union can pose some additional the fang inserted closer to its ventral surface; a obstacles for recognizing palaeognathy as a toothed part of the cheliceral furrow extends pre-neognathous condition. along the full length of this structure; the fang Besides, the trigonotarbid chelicerae with extends backwards horizontally (Fig. 17). In the most pronounced palaeognathy show greater contrast to the plagiaxial type, they were dis- similarity to the mygalomorph variant, rather tinctly flattened prolaterally; the dorsal part of than to the liphistiid and hypochiloid states (Figs the basal sclerite was swollen and slightly 13–15). They also differ notably from the cheli- mound-like [Selden et al., 1991: Pl. 1, fig. 6]. ceral construction known for the most ancestral Finally, it should be emphasized that the highly spiders (see below). Most probably, their nar- orthognathous construction seems to be equally rowed form and the unique axial orientation have suitable as a basis for both the modified orthog- been caused by incapsulating (together with bases nathous and plagiaxial types. of other appendages) into a tight testaceous car- 3 Those isolated chelicerae imprints could belong, in apace provided with a hypertrophied downward- principle, to a fossil arachnid taxon outside the Trigono- inclined clypeus — as shown from the trigono- tarbida. However, the authors of that study suggested that tarbid reconstructions figured by Shear et al. they should be attributed to the fragmentary known trigo- notarbid Gelasinotarbus reticulatus “either by associa- [1987], Dunlop [1996a], etc. In the orthognath tion with carapace fragments or by having characteristic spiders possessing the carapace more open fron- sculpture” [Shear et al., 1987: p. 39]. 356 EUROPEAN ARACHNOLOGY 2003

Principal modifications furrow with two tooth rows aligned throughout the whole length of this structure. Although those characters do not support a The main features of the cheliceral configu- close resemblance between the liphistiid and ration in Heptathela and Liphistius are: the hypochiloid chelicerae, they give ample basis modified (swollen) dorsal lobe of the basal for further cheliceral modifications in the myg- segment; displacement of a fang attachment alomorphs. point closer to its ventral surface; lengthening of the cheliceral furrow. Despite the fact that Mygalomorphae such a construction appears to be very similar to The generalized shape of the chelicerae in the general shape of the chelicerae in the myg- mygalomorphs differs only a little from that alomorphs, lateral surfaces of the basal seg- occurring in the Mesothelae. The most notice- ments in liphistiids appear less convex (i.e., able detail of this shared construction is a swol- more similar to the types observed in the repre- len dorsal lobe of the basal segment; as a result sentatives of other tetrapulmonate orders, Figs the fang attachment point is confined to its 1–3). All the available liphistiid specimens as ventro-apical vertex vs. a primary dorso-apical well as photographs of live mesothele spiders position (Figs 18–25). When compared with show that they never hold their chelicerae ex- cheliceral construction in the liphistiids, the panded laterally, as was figured by Bristowe mygalomorph chelicerae appear in general to [1933: Pl. 5, fig. 9]. Like most mygalomorph be noticeably more highly modified. Additional spiders, the Recent representatives of the Li- modifications which occurred in the Mygalo- phistiidae are burrowers [see Platnick & Sedg- morphae are: (1) enlarged chelicerae in most wick, 1984; Ono, 1999]. members of the Fornicephalae sensu Raven, According to Kraus & Kraus [1993], the li- 1985, achieving their maximum size in the rep- phistiid chelicerae lie close to the plagiaxial mod- resentatives of the Atypidae; (2) short chelicer- el typical for Hypochilus. However, when com- ae of Migidae and some Actinopodidae, and (3) pared these variants show strong evidence that the somewhat diminished chelicerae in web-build- mesothelid state resembles the orthognathous pat- ing mygalomorphs of the families tern occurring in the mygalomorphs, rather than and, in part, Mecicobothriidae and Hexathe- the plagiaxial chelicerae of the Hypochiloidea (cf. lidae. It should be noted that unlike the special- Figs 17–24 and 32). The same relationship is ized burrowing mygalomorphs (but close to the characteristic for the Upper me- Recent liphistiids), the basal cheliceral seg- sothelid Paleothele, whose chelicerae belong ments in the representatives of the latter fami- to the paraxial type [Selden, 1996a4,b]. lies are mostly geniculate, not convex laterally Judging from the description, the chelicerae (as shown in Fig. 5). of Paleothele were rather small, they appear to In diplurids the chelicerae vary in the relat- be smaller than chelicerae in the Recent repre- ed species from rather large to small, even sentatives of the group [see Selden, 1996a: fig. within the same [cf. Raven, 1984a: figs 3b]. Unfortunately, the poor preservation of the 45–51, 124–139; Coyle, 1988: figs 15, 16]. cheliceral parts in these fossils does not give an However, genera wholly known to possess small opportunity for their unequivocal reconstruc- chelicerae with a ‘lesser orthognathy’ (Micro- tion. Even so, in shape they definitely differ hexura Crosby et Bishop, Chilehexops Coyle) from both the phrynid type and the plagiaxial have at the same time some other characters in chelicerae of hypochiloids. The fang position in a clearly derived or reduced form. These in- Paleothele seems to be modified and translo- clude the number of eyes or spines, or the cated clearly closer to the ventral side of the construction of the male mating spur [see Coyle, basal cheliceral segment. Besides, the fossil 1981, 1986]. Within the mecicobothriids the mesothele spider possessed a long cheliceral less developed chelicerae somewhat resembling

4 First described under Eothele Selden; later a new name the plagiaxial construction are known for the was proposed in view of the homonymy [Selden, 2000]. most advanced genus Hexurella Platnick et S.L. Zonstein. Origin and evolution of spider chelicerae 357

Figs 18–21. Chelicerae of liphistiids and primitive mygalomorphs, prolateral view: 18 — Liphistius malayanus (Abraham), Liphistiidae; 19 — Megahexura fulva (Chamberlin), Mecicobothriidae [from Gertsch & Platnick, 1979: fig. 56; modified]; 20 — Atypus sp., Atypidae; 21 — unicolor (Hentz), Antrodiaetidae. Ðèñ. 18–21. Õåëèöåðû ëèôèñòèèä è ïðèìèòèâíûõ ìèãàëîìîðôîâ, ïðîëàòåðàëüíî: 18 — Liphistius malayanus (Abraham), Liphistiidae; 19 — Megahexura fulva (Chamberlin), Mecicobothriidae [ïî Gertsch & Platnick, 1979: fig. 56; èçìåíåíî]; 20 — Atypus sp., Atypidae; 21 — Antrodiaetus unicolor (Hentz), Antrodiaetidae. Gertsch, whereas three, more archaic genera morphs were also burrowing spiders. As for the possess the strictly orthognathous large cheli- ctenizoids (and their closest relatives), the rath- cerae [cf. Gertsch & Platnick, 1979: figs 35–37, er large chelicerae are observed in females, 51–53, 58–60 vs. figs 70–73]. The more ad- known for their higher burrowing activity. vanced position of the first genus was con- Hence, there are grounds to connect the ob- firmed by Raven [1985]. served enlargement of the chelicerae in the The web microstructure in diplurids some- ctenizoid mygalomorph spiders to their further what differs from that known for the araneo- adaptations to their burrowing mode of life. morph spiders [see Palmer, 1985], but at the Separate attention should be paid to the same time is indistinguishable from the burrow short chelicerae of migids and some actinopo- silk lining. Thus, it could be logically assumed dids considered by Kraus & Kraus [1993] as that their ancestors like most Recent mygalo- plagiaxial, and thus closest to the type that 358 EUROPEAN ARACHNOLOGY 2003

Figs 22–25. Chelicerae of higher mygalomorphs, prolateral view: 22 — Anemesia sp., Cyrtaucheniidae; 23 — Cteniza sp., Ctenizidae; 24 — sp., Nemesiidae; 25 — Chaetopelma sp., Theraphosidae. Ðèñ. 22–25. Õåëèöåðû âûñøèõ ìèãàëîìîðôîâ, ïðîëàòåðàëüíî: 22 — Anemesia sp., Cyrtaucheniidae; 23 — Cteniza sp., Ctenizidae; 24 — Damarchus sp., Nemesiidae; 25 — Chaetopelma sp., Theraphosidae. hypothetically could be ancestral for spiders. In contrast, the true archaic mygalomorphs, The problem which appears within the frame- ‘atypoids’, possess well-developed orthogna- work of this hypothesis lies in the fact that the thous chelicerae. Palaeontological evidence mentioned taxa do not belong certainly to a shows that this type was shared by the known number of the most primitive members even Mesozoic mygalomorph spiders [see Eskov & within their own group (Rastelloidina). For in- Zonstein, 1990b; Selden & Gall, 1992; Dunlop, stance, the Migidae are characterized by the 1993]. Taking this into consideration, the pres- secondary loss of the cheliceral rastellum. Be- ence of short chelicerae in some mygalomorph sides, according to Raven’s [1985] opinion, a taxa could be more preferably considered as a very characteristic eye arrangement in the form secondary apomorphic innovation rather than a of a wide trapezium shared by the representa- plesiomorphy. tives of both families, is also a second apomor- Some characters show closer resemblance phic acquisition. between the liphistiid and mygalomorph or- S.L. Zonstein. Origin and evolution of spider chelicerae 359 thognathous chelicerae. Symptomatically, even phistiids (Fig. 26). They only formally resem- when considered in detail, the shape of the ble the short-orthognathous chelicerae of some liphistiid chelicerae appears to be very similar Actinopodidae, Migidae and Idiopidae: in spite to that in the Atypidae and, to a lesser degree, to of the fangs intersected at 90°, the latter possess the configuration of the chelicerae observed in basal cheliceral segments that are flattened pro- two other families of so called ‘atypoid’ myga- laterally and retain the primary longitudinal lomorphs. The grounds for drawing these paral- orientation (Figs 27, 28), typical for mygalo- lels are as follows. morphs. In contrast to the Hypochiloidea, their Firstly, in this case the fang attachment point chelicerae are not directed downward and side- seems not to be displaced completely on the ways [see Platnick & Shadab, 1976; Raven, ventral cheliceral surface (cf. Figs 18–21 vs. Figs 1984b; Goloboff & Platnick, 1987; etc]. How- 22–25). Secondly, both the Liphistiidae and ‘aty- ever, the divergent hypochiloid chelicerae could poid’ mygalomorphs possess a sharp dorsal che- be easily derived from the cheliceral construc- liceral hook separating a swollen dorsal lobe tion of Attercopus, which almost ideally con- from the cheliceral base5, whereas in other myg- forms to the role of the missing link between alomorphs it is developed considerably less, or is two main lineages in the development of spider almost completely absent (see the same figures)6. chelicerae. Furthermore, the hypochiloid con- The last pair of characters was interpreted by dition taken in itself appears to be an intermedi- Eskov & Zonstein [1990b] as an autapomorphy ate that could join the primary neognathous of the ‘atypoids’ and a symplesiomorphy of the type of Attercopus and diaxial chelicerae of the mygalomorphs, respectively. However, the pres- higher araneomorphs. The first and the last ence of a very similar cheliceral hook in the variants are connected through the continuous Liphistiidae either makes their hypothesis doubt- chain of intermediates (as shown in Figs 29–31). ful or at least implies a homoplasy. It should be added that both those types coexist- Raven [1985] gave five apomorphies in or- ed at least since the early Mesozoic because the der to support mygalomorph monophyly: the more advanced diaxial construction existed in completely reduced anterior median spinnerets; the ; later the latter type was widely the subsegmented basal joint of the posterior distributed within the Mesozoic araneomorphs lateral spinnerets; the anterior lateral spinnerets [see Eskov, 1984, 1987; Selden, 1990, 2001; are much smaller than the posterior lateral ones; Selden & Dunlop, 1998; Selden et al., 1999]. both the transverse duct of the posterior lungs It is most likely that the plagiaxial model and the apodemes are reduced entirely; the re- underlay the development of all the types of duced number of the cardiomeres — four instead diaxiality. The so-called paralabidognathous of five in the liphistiids. The peculiar form of the chelicerae of the Filistatidae derived from the mygalomorph chelicerae representing a further orthognathous state by Eskov & Zonstein modification of the neognathous mesothele type [1990a], showed on more careful examination a could be added to these apomorphies irrespec- doubtless resemblance to the cheliceral con- tive of the viewpoints considered here. struction of the hypochilids (cf. Figs 32 and 33). Araneomorphae This fact contradicts the concept of Eskov & As noted above, the plagiaxial chelicerae of Zonstein [1990a] supporting the separate evolu- Hypochilus and their affinities differ noticeably tion of the labidognathous filistatid chelicerae. from the orthognathous construction in the li- Evolution 5 This cheliceral type was also figured by Coyle [1968, 1971, 1974] for Antrodiaetidae and by Gertsch & Possible evolutionary factors Platnick [1979, 1980] for Mecicobothriidae and Atypi- dae, respectively. The similar chelicerae of fossil Creta- It is most likely that an improved web- ceous mecicobothriids and antrodiaetids were described building activity connected with feeding on and figured by Eskov & Zonstein [1990b]. flying and saltatory was the main factor 6 The few exceptions noted in passing by Eskov & Zonstein [Op.cit.: fig. 26d–f] do not contradict the general that induced the origin and further development mode. of labidognathy in spiders [Starobogatov, 1985; 360 EUROPEAN ARACHNOLOGY 2003

Figs 26–31. Principal cheliceral constructions in spiders, ventral view: 26 — Liphistius sp. [Platnick & Gertsch, 1976: fig. 5, modified]; 27 — mygalomorph Neocteniza osa Platnick et Shadab, female [Platnick & Shadab, 1976: fig. 7]; 28 — Neocteniza fantastica Platnick et Shadab, male [Op.cit.: fig. 8]; 29 — Hypochilidae; 30 — some Ctenidae; 31 — majority of araneomorph spiders (as a typical representative, the diguetid Segestrioides bicolor Keyserling is shown according to Platnick [1989: fig. 8]. Ðèñ. 26–31. Îñíîâíûå òèïû êîíñòðóêöèè õåëèöåð ó ïàóêîâ, âåíòðàëüíî: 26 — Liphistius sp. [Platnick & Gertsch, 1976: fig. 5, èçìåíåíî]; 27 — mygalomorph Neocteniza osa Platnick et Shadab, ñàìêà [Platnick & Shadab, 1976: fig. 7]; 28 — Neocteniza fantastica Platnick et Shadab, ñàìåö [Op.cit.: fig. 8]; 29 — Hypochilidae; 30 — íåêîòîðûå òàêñîíû Ctenidae; 31 — áîëüøèíñòâî àðàíåîìîðôíûõ ïàóêîâ (â êà÷åñòâå òèïè÷íîãî ïðåäñòàâèòåëÿ ýòîé ãðóïïû ïðèâåäåíà èëëþñòðàöèÿ Segestrioides bicolor Keyserling, Diguetidae èç Platnick [1989: fig. 8]. S.L. Zonstein. Origin and evolution of spider chelicerae 361

Figs 32–33. Chelicerae of primitive araneomorph spiders, prolateral view: 32 — Hypochilus coylei Platnick, Hypochilidae; 33 — insidiatrix (Forskål), Filistatidae. Ðèñ. 32–33. Õåëèöåðû ïðèìèòèâíûõ àðàíåîìîðôîâ, ïðîëàòåðàëüíî: 32 — Hypochilus coylei Platnick, Hypochilidae; 33 — Filistata insidiatrix (Forskål), Filistatidae. Kraus & Kraus, 1993]. Due to the improvement According to Kraus & Kraus [1993], the of spider webs, efficient prey fixation in the development of orthognathy could be connect- araneomorphs became more dependent on uti- ed to a specific method of prey capture, when lizing the web. This very reason, perhaps, caused the prey is captured by spiders near the ground the further evolution of the diminished semi- to surface. This hypothesis appears to be better fully diaxial chelicerae characteristic for the founded. Most taxa belonging to the Trigono- Araneomorphae. However, the factors which tarbida actually possessed downward-directed caused the appearance of the highly-modified orthognath chelicerae [see Shear et al., 1987: chelicerae in the ancestors of both the Me- fig. 54]. It should be noted, however, that this sothelae and the Mygalomorphae remain ob- type of cheliceral modification was connected scure and somewhat controversial. probably to the development of the ‘incapsulat- Starobogatov [1985] connected the devel- ed’ state of the cephalothorax provided with a opment of large orthognathous chelicerae with hypertrophied clypeus [Op.cit.: fig. 55]. Obser- feeding by the orthognath spider ancestors on vations on the hunting mygalomorph spiders crustaceans and other armored . He show that they do not attempt to clutch the suggested that “the paraxial chelicerae are un- captured prey to the ground surface. In contrast, doubtedly more convenient when the prey should they often try to raise the all too active victim be shreded (because either it is too large, or it is upwards, thus decreasing its ability to resist. defended by a shell); otherwise, the diaxial None of the other arachnids possess paraxial construction is more opportune when the prey is chelicerae of similar shape. Hence, to explain the captured and squeezed close to the mouth. It factors which caused the appearance of these looks reasonable that the paraxial chelicerae modifications, one needs to focus rather on those have been yet formed in those spiders, for which features which are shared at least by most of the the insects were not a principal food source, but orthognath spiders; they also should be absent diaxial chelicerae were aquired in view of a or at least underdeveloped within other orthog- transition to feeding exclusively on the insects” nath arachnids. The only character correspond- [Op.cit.: p. 8]. Unfortunately, this assumption ing to both above conditions is probably the cannot be either confirmed or disproved. None burrowing activity common to the majority of of the more realible characters supporting the spiders possessing the paraxial chelicerae. given trophic specialization of the orthognath Compared to the labidognathous chelicer- spider ancestors can be observed in the palae- ae, the orthognathous construction shows lower ontological material. efficiency, as was noted by Kaestner [1956]. 362 EUROPEAN ARACHNOLOGY 2003

This inefficiency could indicate that the shape But the further evolution of the spider chelicer- of the spider orthognathous chelicerae actually ae suggested by this hypothesis looks improba- represents an evolutionary compromise result- ble since it considers the distinctly modified ing from two competing functions of the same chelicerae of the mygalomorphs as a precursor structure caused by burrowing and hunting ac- for the labidognathous type. One must agree tivities. From this point of view, the configura- with the opinion of Kraus & Kraus [1993] re- tion of the spider paraxial chelicerae seems to garding the mygalomorph and araneomorph che- be rather suitable to unite both these functions. licerae as developed separately from each other. The lower diversity of cheliceral forms with- In principle, the peculiar form of the latter could in the orthognath spiders probably testifies in be explained by a reverse of the basic phrynid favour of the assumption that they became bur- configuration though redirected out- and down- rowers in the early stages of their evolution, i.e., wards. But contrary to that, it could represent their adaptive radiation took place after the or- only a quasi-similar innovation developed through thognath spider chelicerae had been modified. modification of the plagiaxial type. The presence of forward-directed neognathous In its part, the latter hypothesis does not yet chelicerae in ancestral orthognath spiders could contradict the traditional spider classification be considered the necessary preadaptation for because it places the chelicerae of the liphisti- the specific but narrow radiation of these groups. ids into the plagiaxial type; thus it assumes that the origin of the orthognathous and labidogna- Some evolutionary aspects thous variants took place when the suborder It is considered that the chelicerae of the was split into the corresponding tetrapulmonates arose from the three-segment- infraorders. However, as stated above, the che- ed chelate type of the lower arachnids [Dunlop, licerae of the Recent Liphistiidae cannot be 1997]; such a generalized precursor is shown in referred to the same type that includes the pla- Fig. 34a. The cheliceral structures of Amblypy- giaxial construction of the Hypochilidae. Judg- gi and Uropygi s.lat., where the majority of ing from the fossils of Paleothele, the cheliceral characters appear to have an intermediate posi- variant of Hypochilus and those of fossil Paleo- tion between two opposite states found in spi- zoic mesothele spiders were also dissimilar. ders and most primitive arachnids, are regarded The cheliceral configuration found in the here as transitional variants that could represent Devonian Attercopus (Fig. 34f) represents possi- a base for further cheliceral evolution (Fig. bly the only version that appears suitable to be a 34b,c). Contrary to Dunlop [1997], in this study prototype for all the Recent spider groups. Fol- the trigonotarbid chelicerae are supposed to lowing this assumption, one can describe, con- have been further modified to represent a direct sistently and without reversals, the further evolu- precursor of the spider chelicerae (see above). tion of chelicerae in the main groups of the It should be added that in orthognath spiders the Araneae: the reinforcement of paraxial chelicer- fang insertion point seems to be displaced ven- ae in mesothele spiders (Fig. 34g) and their trad through disproportional development of supplementary modification in the mygalomor- upper and lower portions of the basal cheliceral phs (Fig. 34h,i); the origin of paraxial chelicerae segment. Seemingly, in trigonotarbids a similar in the hypochiloid spiders (Fig. 34j) and their configuration was achieved mainly because their gradual modification into the diaxial cheliceral axial orientation had been changed (Fig. 34d,e). construction of the more advanced araneomorph Their resemblance to the cheliceral configura- spiders (Fig. 34k). Although for Attercopus too tion in spiders could represent homoplasy. little is known to conclude anything sufficient The traditional viewpoint appears to be quite concerning its relationships, this taxon was a correct in treating the mygalomorph chelicerae derivative of the primary spider group that gave as derived from the orthognathous mesothele rise to the liphistiids [both taxa share similarly type. In this case, the cheliceral constructions shaped spigots; Selden et al., 1991] and proba- which occurred in the lower mygalomorphs bly to all other spiders. Hence, based on the allow us to consider them the halfway variants. above-mentioned data, the mygalomorph and S.L. Zonstein. Origin and evolution of spider chelicerae 363

Fig. 34. Presumed evolution of chelicerae in main lineages of tetrapulmonate arachnids (with comments in the text): A — lower arachnids (Scorpiones — according to Stockwell, 1996, modified); B — schizomid Stenochrus [Tourinho & Kury, 1999, fig. 5]; C — whip spider Charinus [Giupponi & Kury, 2002: fig. 2]; D — some Trigonotarbida [Dunlop, 1997: fig. 4a]; E — trigonotarbid Gelasinotarbus reticulatus [Shear et al., 1987, fig. 68]; F — Attercopus [Selden et al., 1991: fig. 4c]; G — Liphistius; H — mecicobothriid Megahexura [Gertsch & Platnick, 1979: fig. 56, modified]; I — Hypochilus; J — nemesiid [Loksa, 1966: fig. 1h, modified]; K — majority of Araneomorphae. Fig. 34. Ïðåäïîëàãàåìàÿ ýâîëþöèÿ õåëèöåð â ãëàâíûõ ãðóïïàõ (êîììåíòàðèè ïðèâåäåíû â òåêñòå): A — íèçøèå àðàõíèäû (Scorpiones — ïî Stockwell, 1996, èçìåíåíî); B — ñõèçîìèä Stenochrus [Tourinho & Kury, 1999: fig. 5]; C — ôðèí Charinus [Giupponi & Kury, 2002: fig. 2]; D — íåêîòîðûå òàêñîíû Trigonotarbida [Dunlop, 1997: fig. 4a]; E — òðèãîíîòàðáèä Gelasinotarbus reticulatus [Shear et al., 1987: fig. 68]; F — Attercopus [Selden et al., 1991: fig. 4c]; G — Liphistius; H — ìåöèêîáîòðèèä Megahexura [Gertsch & Platnick, 1979: fig. 56, èçìåíåíî]; I — Hypochilus; J — íåìåçèèä Nemesia [Loksa, 1966: fig. 1h, èçìåíåíî]; K — áîëüøèíñòâî òàêñîíîâ Araneomorphae. 364 EUROPEAN ARACHNOLOGY 2003

Table. Summary of character patterns of the variants considered. Òàáëèöà. Ñóììèðîâàííûå ïàòòåðíû ïðèçíàêîâ âñåõ ðàññìîòðåííûõ âàðèàíòîâ.

Character Traditional concept Kraus & Kraus [1993] This study Primary cheliceral paraxial chelicerae plagiaxial chelicerae paraxial chelicerae type within the Araneae Chelicerae of the ancestral or close arose independently ancestral or close Amblypygi to basic type from the Araneae to basic type Chelicerae of the plesiomorphically close to plagiaxial type orthognathous but Mesothelae orthognathous modified (neognathous) Chelicerae of the plesiomorphically represent further represent further Mygalomorphae orthognathous modification of modification of the mesothele type the mesothele type Chelicerae of the arose through arose independently from arose independently Araneomorphae orthognathous type orthognathous type from orthognathous type Fang insertion point [not specified] unmodified displaced ventrad in the Mesothelae and the Mygalomorphae Axial orientation unmodified modified partially modified partially in the Mygalomorphae from plagiaxial from paraxial to paraxial type to plagiaxial type Plesiomorphic state of long furrow and fang short furrow and fang long furrow and fang cheliceral furrow and fang in the Mygalomorphae Cheliceral modifications peculiarities of feeding peculiarities burrowing activity in orthognath spiders [Starobogatov, 1985] of prey capture result from: araneomorph spiders might not be sister groups, cerae is questioned. This hypothesis seems to though it is the conventional hypothesis. have overlooked the fact that the described On the other hand, the liphistiomorph and variants of chelicerae both in the Mesothelae mygalomorph spiders could have acquired their and in the Mygalomorphae differ considerably cheliceral type independently from each other. It from the hypothesized ancestral state. should be noted that both groups are represented 2. The new approach first shown by Kraus mainly by burrowing forms with a restricted area & Kraus [1993] appears considerably more of potential modifications. Therefore, such mod- suitable to explain parsimonously the origin ifications should have been substantially nar- and evolution of labidognathous chelicerae. Im- rower in scope than those seen in the araneomor- mediately, the hypothesis faces difficulties when phs. Such a concept appears to be more parsimo- it derives the orthognathous chelicerae from the nious since it eliminates the problem of multiple hypothetical intermediate type, common with parallel reductions in the Mygalomorphae and the labidognaths. The facts contradicting this the Araneomorphae required by the first hypoth- hypothesis are as follows: (1) the most archaic esis. However, the great similarity existing be- mygalomorphs possess concurrently well-de- tween cheliceral configurations in the liphistiids veloped orthognathous chelicerae; (2) chelicer- and the most archaic mygalomorphs in the frame- ae of the Liphistiidae also differ noticeably work of this variant remains unexplained. from the assumed ancestral type; (3) the latter was not found in the pre-Cenozoic mesothele Conclusions and mygalomorph fossils discovered in the last 15 years; like their Recent relatives, these forms 1. The traditional concept treating the cheli- also possessed paraxial chelicerae. ceral construction in the orthognathous spiders 3. In the hypothesis proposed here the che- as an initial state of the labidognathous cheli- licerae both of the Liphistiidae and the Mygalo- S.L. Zonstein. Origin and evolution of spider chelicerae 365 morphae can be treated as plesiomorphic by References their position and axial orientation, but as apo- morphic by their configuration; the latter seems Bristowe W.S. 1933. The liphistiid spiders. With an ap- to be an evolutionary compromise caused by pendix on their internal by J. Millot // Proc. burrowing activity. Whether the shared type in Zool. Soc. Lond. Vol.103. P.1016–1057. the above groups is a synapomorphy or a ho- Coyle F.A. 1968. The mygalomorph spider genus Aty- poides (Araneae, Antrodiaetidae) // Psyche. Vol.75. moplasy remains uncertain although the cheli- P.157–194. ceral construction of archaic mygalomorphs Coyle F.A. 1971. Systematics and natural history of the shows some resemblance to that of liphistiids. mygalomorph spider genus Antrodiaetus and related In contrast, the cheliceral position in the Arane- genera (Araneae, Antrodiaetidae) // Bull. Mus. Comp. Zool. Harv. Vol. 141. P.269–402. omorphae seems to be a definite autapomor- Coyle F.A. 1974. Systematics of the trapdoor spider genus phy, yet their configuration being much closer Aliatypus (Araneae, Antrodiaetidae) // Psyche. Vol. (as compared to the above mentioned groups) 81. P.431–500. to that in the other tetrapulmonate orders should Coyle F.A. 1981. The mygalomorph spider genus Micro- hexura (Araneae, Dipluridae) // Bull. Am. Mus. Nat. be considered either a symplesiomorphy or a Hist. Vol. 170. P.64–75. reversal. The distinctions occurred between the Coyle F.A. 1986. Chilehexops, a new funnelweb mygalo- traditional approach, Kraus & Kraus hypothe- morph spider genus from Chile (Araneae, Dipluridae) sis of 1993, and the new version proposed in // Am. Mus. Novit. No.2860. P.1–10. Coyle F.A. 1988. A revision of the American funnel-web this study, are summarized in the Table. mygalomorph spider genus Euagrus (Araneae, Diplu- 4. The observed data generally support the ridae) // Bull. Am. Mus. Nat. Hist. Vol.187. P.203– Kraus & Kraus [1993] hypothesis, namely that 292. the two main cheliceral types arose indepen- Dunlop J.A. 1993. A review of fossil mygalomorphs // Mygalomorph. Vol.1. P.1–17. dently of one another. Unlike the cheliceral Dunlop J.A. 1994. Filtration mechanisms in the mouth- construction which occurred in other orders of parts of tetrapulmonate arachnids // Bull. Br. Arach- the Tetrapulmonata, paraxial chelicerae of the nol. Soc. Vol.9. No.8. P.605–614. Recent orthognath spiders are modified to such Dunlop J.A. 1996a. A trigonotarbid arachnid from the Upper of Shropshire // Palaeontology. Vol.39. an extent that they cannot represent the basic P.605–614. construction for the Araneomorphae. Dunlop J.A. 1996b. Systematics of the fossil arachnids // Rev. Suisse Zool. (Vol. Hors Série 2). P.173–184. ACKNOWLEDGEMENTS. I am cordially grate- Dunlop J.A. 1996c. Evidence for a sister group relation- ful to Dr. N. Platnick (American Museum of Natural ship between Ricinulei and Trigonotarbida // Bull. Br. Arachnol. Soc. Vol.10. No.6. P.193–204. History, New York), Dr. F. Coyle (Western Caroli- Dunlop J.A. 1997. Paleozoic arachnids and their signifi- na University, Cullowhee NC) and Dr. K. Eskov cance for arachnid phylogeny // ¯abka M. (ed.). Proc. (Paleontological Institute RAS, Moscow) for pro- 16th Europ. Coll. Arachnol., Siedlce, Poland. P.65–82. viding material, and also to Prof. Dr. O. Kraus Eskov K.Yu. 1984. A new fossil spider family from the (Zoologisches Institut und Zoologisches Museum, of Transbaikalia (Araneae: ) // Universität Hamburg), Dr. W. Shear (Department of Neues Jahrbuch Geol. Paläontol., Monatshefte. Hft. Entomology, Hampden-Sydney College), Prof. Dr. 1984. S.645–653. P. Weygoldt (Albert-Ludwigs-Universität, Freiburg), Eskov K.Yu. 1987. A new archaeid spider (Chelicerata: Dr. Yu. Marusik (Institute of Biological Problems Araneae) from the Jurassic of Kazakhstan, with notes on the so-called ‘Gondwanan’ ranges of recent taxa of the North, Magadan), Dr. P. Selden (Department // Neues Jahrbuch Geol. Paläontol. Abhandl. Bd.175. of Earth Sciences, University of Manchester), Dr. J. Hft.1. S.81–106. Dunlop (Institut für Systematische Zoologie, Hum- Eskov K.Yu. & Zonstein S.L. 1990a. A new classification boldt-Universität, Berlin) and Mr. A. Gromov (In- for the order Araneida (Arachnida: Chelicerata) // stitute of Zoology KazAS, Almaty) for help with Acta Zool. Fenn. Vol.190. P.129–137. obtaining literature. Dr. D. Logunov (The Manches- Eskov K.Yu. & Zonstein S.L. 1990b. First Mesozoic myg- ter Museum, University of Manchester) helped me alomorph spiders from lower of Siberia and with submitting the manuscript and Dr. V. Fet (Mar- Mongolia, with notes on the system and evolution of the infraorder Mygalomorphae // Neues Jahrbuch Geol. shall University, Huntington WV) and Dr. A. Fre- Paläontol. Abhandl. Bd.178. Hft.3. S.325–368. idberg (Department of Entomology, Tel-Aviv Uni- Gertsch W.J. & Platnick N.I. 1979. A revision of the spider versity) kindly checked my English. I am grateful to family Mecicobothriidae (Araneae, Mygalomorphae) two anonymous reviewers for manuscript comments. // Am. Mus. Novit. No.2687. P.1–32. 366 EUROPEAN ARACHNOLOGY 2003

Gertsch W.J. & Platnick N.I. 1980. A revision of the Raven R.J. 1984b. Systematics and biogeography of the American spiders of the family Atypidae (Araneae, mygalomorph spider family Migidae (Araneae) in Mygalomorphae) // Am. Mus. Novit. No.2707. P.1–39. Australia // Aust. J. Zool. Vol.32 P.379–390. Giribet G., Edgecombe G.D., Wheeler W.C. & Babbitt C. Raven R.J. 1985. The spider infraorder Mygalomorphae 2002. Phylogeny and systematic position of Opil- (Araneae): Cladistics and systematics // Bull. Amer. iones: a combined analysis of chelicerate relation- Mus. Nat. Hist. Vol.182. P.1–180. ships using morphological and molecular data // Cla- Selden P.A. 1990. Lower Cretaceous spiders from the distics. Vol.18. P.5–70. Sierra de Montsech, north-east Spain // Palaeontolo- Giupponi A.P. & Kury A.B. 2002. A new species of gy. Vol.33. P.257–285. Charinus Simon, 1892 from northeastern Brazil Selden P.A. 1996a. First fossil mesothele spider from the (Arachnida, Amblypygi. Charinidae) // Bol. Mus. Carboniferous of France // Rev. Suisse Zool. (Vol. Nac. Zool. Rio. No.477. P.1–7. Hors Série 2). P.585–596. Goloboff P.A. & Platnick N.I. 1987. A review of the Chilean Selden P.A. 1996b. Fossil mesothele spiders // Nature. spiders of the superfamily Migoidea (Araneae, Myga- Vol.379. P.498–499. lomorphae) // Am. Mus. Novit. No.2888. P.1–15. Selden P.A. 2000. Paleothele, replacement name for the Grasshoff M. 1978. A model of the evolution of the main fossil mesothele spider Eothele Selden non Rowell // chelicerate groups // Symp. Zool. Soc. Lond. No.42. Bull. Br. Arachnol. Soc. Vol.11. P.272. P.273–284. Selden P.A. 2001. spiders from the Isle of Wight Harvey M.C. 2002. The neglected cousins: what we do with preserved respiratory structures // Palaeontology. know about the smaller arachnid orders? // J. Arach- Vol.39. P.583–604. nol. Vol.30. P.357–372. Selden P.A. & Dunlop J.A. 1998. Fossil taxa and relation- Harvey M.C. 2003. Catalog of the smaller arachnid orders ships of chelicerates // Edgecombe G.D. (ed.). Arthro- of the world. Amblypygi, Uropygi, Schizomida, Pal- pod fossils and phylogeny. New-York: Columbia Univ. pigradi, Ricinulei and . CSIRO Publ. 400 p. Press. P.303–331. Kaestner A. 1956. Lehrbuch der speziellen Zoologie. Wir- Selden P.A. & Gall J.-C. 1992. A Triassic mygalomorph bellose. Bd.1. Hft.5. S.485–658. spider from the northern Vosges, France // Palaeonto- Kraus O. & Kraus M. 1993. Divergent transformation of logy. Vol.35. P.211–235. chelicerae and original arrangement of eyes in spiders Selden P.A., Shear W.A. & Bonamo P.M. 1991. A spider and other arachnids from the Devonian of New York, (Arachnida, Araneae) // Mem. Queensland Mus. and reinterpretations of Devonian Araneae // Palaeon- Vol.33. No.2. P.579–584. tology. Vol.34. P.241–281. Loksa I. 1966. Nemesia pannonica O. Herman (Araneae, Selden P.A., Anderson J.M., Anderson H.M. & Fraser Ctenizidae) // Ann. Univ. Sci. Budap. Sect. Biol. N.C. 1999. Fossil araneomorph spiders from the Tri- Vol.8. P.155–171. assic of South Africa and Virginia // J. Arachnol. Mikhailov K.G. 1992. Third Eurasian Arachnological Vol.27. P.401–414. Conference // Arthropoda Sel. Vol.1. No.3. P.99–100. Shear W.A., Selden P.A., Rolfe W.D.I., Bonano P.M. & Miller J.A. 2003. Assessing progress in systematics with Grierson J.D. 1987. New terrestrial arachnids from the continuous jackknife function analysis // Syst. Biol. Devonian of Gilboa, New York (Arachnida, Trigono- Vol.52. No.1. P.55–65. tarbida) // Am. Mus. Novit. No.2901. P.1–74. Ono H. 1999. Spiders of the genus Heptathela (Araneae, Shultz J.W. 1990. Evolutionary morphology and phylog- Liphistiidae) from Vietnam, with notes on their natu- eny of Araneida // Cladistics. Vol.6. P.1–38. ral history // J. Arachnol. Vol.27. P.37–43. Shultz J.W. 1999. Muscular anatomy of a whip spider, Palmer J.M. 1985. The silk and silk production system of the Phrynus longipes (Pocock) (Arachnida, Amblypygi), funnel-web mygalomorph spider genus Euagrus (Ara- and its evolutionary significance // Zool. J. Linn. Soc. neae, Dipluridae) // J. Morph. Vol.186. P.195–207. Vol.126. P.81–116. Platnick N.I. 1989. A revision of the spider genus Segestri- Starobogatov Ya.I. 1985. [The taxonomic position and oides (Araneae, Diguetidae) // Am. Mus. Novit. system of the spiders] // Ovtsharenko V.I. (ed.). Fauna No.2940. P.1–9. i ekologiya paukov SSSR. Trudy Zool. Inst. AN SSSR. Platnick N.I. & Gertsch W.J. 1976. The suborders of Vol.139. P.4–16 [in Russian, with English summary]. spiders: a cladistic analysis (Arachnida, Araneae) // Stockwell S. 1996. terminology and morpholo- Am. Mus. Novit. No.2607. P.1–15. gy, online at http://wrbu.si.edu/www/stockwell/mor- Platnick N.I. & Sedgwick W.C. 1984. A revision of the phology/morphology.html spider genus Liphistius (Araneae, Mesothelae) // Am. Tourinho A.L.M. & Kury A.B. 1999. The southernmost Mus. Novit. No.2781. P.1–31. record of Schizomida in South America, first records of Platnick N.I. & Shadab M.U. 1976. A revision of the Schizomida or Rio de Janeiro and of Stenochrus Cham- mygalomorph spider genus Neocteniza (Araneae, berlin, 1922 for Brazil (Arachnida, Schizomida, Hub- Actinopodidae) // Am. Mus. Novit. No.2603. P.1–19. bardiidae) // Bol. Mus. Nac. Zool. Rio. No.405. P.1–6. Pocock R.I. 1893. On some point in the morphology of the Weygoldt P. & Paulus H.F. 1979. Untersuchungen zur Mor- Arachnida (s.str.) with some notes on the classification of phologie, Taxonomie und Phylogenie der Chelicerata. II. the group // Ann. Mag. Nat. Hist. Ser.6. No.11. P.1–19. Cladogramme und die Entfaltung der Chelicerata // Z. Raven R.J. 1984a. Systematics of the Australian curtain- Zool. Syst. Evolutionsforschung. Bd.17. S.177–200. web spiders (Ischnothelinae: Dipluridae: Chelicerata) Wheeler W.C. & Hayashi C.Y. 1998. The phylogeny of extant // Aust. J. Zool. Supplry. Ser. Vol.93 P.1–102. chelicerate orders // Cladistics. Vol.14. P.173–192.