Revision of the enigmatic Southeast Asian spider genus Savarna ( Araneae , Pholcidae )

The genus Savarna Huber, 2005 was previously one of the most poorly known Pholcinae genera. Less than 20 specimens (representing four nominal species) were available worldwide; nothing was known about ultrastructure, natural history, or relationships. We present the first SEM data, supporting the position of the genus in Pholcinae outside the Pholcus group of genera and weakly suggesting a closer relationship with the genera Khorata Huber, 2005, Spermophorides Wunderlich, 1992, and two undescribed species of unknown affinity from Borneo. We provide the first data about microhabitat, web structure, and reaction to disturbance. We clarify the type locality of Savarna tessellata (Simon, 1901) (“Jalor, Biserat”) and describe topotypical material. We describe the previously unknown male of Spermophora miser Bristowe, 1952 and transfer the species (that was previously considered incertae sedis) to Savarna as Savarna miser (Bristowe, 1952) comb. nov. Savarna baso (Roewer, 1963) is newly synonymized with S. miser. We describe the most northern species in the genus, Savarna kaeo sp. nov., and provide amendments to the descriptions of all previously described species.


Introduction
The genus Savarna Huber, 2005 was one of several genera created to accommodate species previously misplaced in Spermophora Hentz, 1841 (Huber 2005).Originally, the genus included three species, Savarna tessellata (Simon, 1901) and S. thaleban Huber, 2005 from southern Thailand, and S. baso (Roewer, 1963) from Sumatra.At that time, only four male and nine female specimens were known.The females of S. tessellata and S. baso could not be fully redescribed, either because the epigynum of the only female specimen was distorted (S. tessellata) or because the abdomen of the only female specimen was missing (S. baso).Nothing was known about the biology of any species.No data were available about ultrastructure.No material was available for molecular work.Nothing was known about relationships within the genus or with other genera in the subfamily Pholcinae.

Monophyly and relationships
Previously, the monophyly of Savarna was supported mainly by two unique morphological characters, the modifications of the male chelicerae and of the male palpal trochanter (Huber 2005).Preliminary analyses of molecular data of all known species (A. Valdez-Mondragón, B.A. Huber, D. Dimitrov, unpubl. data) strongly support the monophyly of Savarna.
The single ventral row of comb-hairs on the fourth tarsus places the genus in Pholcinae (cf.Huber & Fleckenstein 2008).The morphology of the comb-hairs ('Belisana-type' sensu Huber & Fleckenstein 2008) suggests the genus is not part of the Pholcus group of genera.Preliminary analyses of molecular data strongly support this 'basal' position in Pholcinae.Detailed relationships with other genera among 'basal' Pholcinae remain obscure.The absence of epiandrous spigots is rare in Pholcinae and is otherwise known only in Khorata Huber, 2005 (B.A. Huber, unpubl.data on three species), in Spermophorides Wunderlich, 1992 (B.A. Huber, unpubl.data on one species), and in two undescribed species of unknown affinity from Borneo (B.A. Huber, unpubl. data).Our preliminary molecular analyses do weakly support a close relationship among some of these taxa.

Natural history
Even though Savarna spiders also occur outside caves, all known localities are closely associated with limestone caves.Both inside and outside the caves, the spiders build domed webs close to the ground with at least one side of the web attached to the rock surface.Web diameter ranges from about 10 to 25 cm.When disturbed, the spiders vibrate; if disturbance continues or is strong, they drop to the ground and remain motionless for a while (cf.Fig. 7).Egg-sacs are covered by a sparse layer of barely visible silk and contain about 20-30 eggs (cf.Fig. 13).For data on individual species, see descriptions below.

Composition
The genus now includes five species, all of which are treated below.

Distribution
Savarna appears to be restricted to southern Thailand, mainland Malaysia, and Sumatra (Fig. 1; but see Addendum).

Etymology
The species name is derived from the type locality; noun in apposition.Color.Carapace pale ochre, with wide dark lateral margins; ocular area and clypeus also dark brown; sternum black; legs light ochre, with darker rings on femora (subdistally) and tibiae (proximally); abdomen gray, with dark subcuticular marks lying above deeper white marks, with distinct ventral pattern consisting of three interconnected black marks.

Material examined
Body.Habitus as in Fig. 2; ocular area elevated, each triad on short hump directed toward lateral; carapace with deep median furrow (Fig. 19); clypeus with pair of rounded processes at rim provided with strong and long hairs ; sternum wider than long (0.72 / 0.48), unmodified.Chelicerae as in Fig. 16, with pair of lateral processes directed slightly toward posterior; without stridulatory ridges.ALS as in female (cf.Fig. 24).Gonopore without epiandrous spigots (Fig. 26).

Male (variation)
Numbers of spines on legs variable, femur 3 spines absent in most males; small males with fewer and thinner spines.Tibia 1 in 16 other males: 4.9-6.4(mean 5.6).

Natural history
Spiders were found both inside caves (twilight zone) and outside caves if large rocks provided sufficient shade (Tham Kaeo; Tham Phraya Nakhon).When disturbed, the spiders first vibrated, and then dropped to the ground, remaining motionless.

Amendments to original description
Male clypeus without processes but with strong hair brushes not present in female .Male gonopore without epiandrous spigots (Fig. 52).Male and female ALS with only two spigots each (Fig. 53).In the palp illustrated in the original description (Wongprom & Wiwatwitaya 2015: fig.1a) the bulb is rotated about 180° from its natural position.In the natural position, the long pointed process is directed in the opposite direction and the proximal bulbal sclerite is visible in prolateral view.Spines on male legs sometimes absent, sometimes present (each femur 1 with two ventral rows of up to ~ 25 spines each); tibia 1 L/d: 53; prolateral trichobothrium absent on tibia 1, present on other tibiae; male and female tarsus 4 with single row of ventral comb-hairs (Fig. 51).Tibia 1 in 10 males: 5.7-6.8(mean: 6.3); in 5 females: 4.7-5.6 (mean 5.3).

Natural history
Spiders were found both inside and outside the cave, in domed webs of ~ 20 cm diameter.In the cave, they only occurred in the twilight zone.Outside the cave, they were found to be most abundant in the small forest above the cave.When lightly disturbed, the spiders vibrated in their webs; when disturbed more strongly, spiders dropped to the ground, remaining motionless (cf.Fig. 7).

Amendments to original description
Male clypeus with pair of small lateral processes at rim (Fig. 55).Male gonopore without epiandrous spigots.Male and female ALS with only two spigots each (Fig. 61).In the palp illustrated in the original description (Huber 2005: figs 131-132), the bulb is rotated about 180° from its natural position.In the natural position, the long bulbal process is directed in the opposite direction and the proximal bulbal sclerite is visible in prolateral view.All males seen without leg spines; prolateral trichobothrium absent on tibia 1, present on other tibiae; male and female tarsus 4 with single row of ventral comb-hairs (Fig. 62).Tibia 1 in 8 males: 4.5-5.4(mean: 5.0); in 9 females: 4.2-4.9(mean 4.5).

Natural history
All specimens were found in the forest outside a small cave, in domed webs among rocks close to the ground.Spiders were abundant but fled very rapidly, dropping from the web to the ground and becoming essentially invisible.

Notes on type material and type locality
The three specimens (1♂, 1♀, 1 juv.) in MNHN redescribed in Huber (2005) may or may not include the female described by Simon (1901).Simon's handwritten label just says "12185 Sp. tessellata E.S. Pen.Malayana (C.M)".In any case, we are now confident that these specimens, as well as the new specimens collected at Wat Kuhapimuk (see below), are in fact conspecific with Simon's described specimen(s).All details of Simon's description fit the newly collected females, and even the distinctive shape of the epigynum ("…in medio depressa, … utrinque oblique truncata, postice leviter prominula"; Simon 1901: 50) exactly fits the new specimens.
"Jalor, Biserat" was previously erroneously thought to be in Malaysia (Huber 2005).Instead, several lines of evidence suggest that this is in Yala Province (=Jalor in the Pattani Malay language) in southern Thailand.We actually suggest that the type locality is precisely what is today Wat Kuhapimuk, near Yala City.First, Laidlaw's (1900) brief account of the itinerary of the Skeat Expedition shows that "Biserat, a village in a small state called Jalor" was visited on the way between Patani and the Kelantan River in northeastern Malaysia.Yala City is located exactly between Patani (=Pattani) and the Kelantan River.Second, Serat (Pattani Malay language) equals Sap (Thai language), suggesting that the village Ban Tha Sap (Thai for 'village -riverbank -Sap'), located just 2 km from Wat Kuhapimuk, might be identical to Biserat.Finally, another pholcid species described by Simon (1901) in the same publication (Pholcus diopsis) has the type locality "Gua Glap ("Dark Cave"), Biserat, Jalor" (Simon 1901: 50).We found Pholcus diopsis in the cave Tham Meud at Wat Kuhapimuk.Thus, we assume that what is today called Tham Meud (Thai = Dark Cave) at Wat Kuhapimuk is identical to Simon's Gua Glap (Malay = Dark Cave).Unlike Pholcus diopsis, Savarna tessellata occurs both inside and outside the cave.The original material was possibly collected outside, explaining the absence of a mention of the cave in the original description.
Color.Carapace pale ochre, without dark lateral margins; ocular area and clypeus dark brown; sternum black; legs light brown, with darker rings on femora (subdistally) and tibiae (proximally and subdistally); abdomen gray, with dark and white marks, large whitish mark above spinnerets (Fig. 11), with distinct ventral pattern consisting of three interconnected black marks.

Natural history
Both in and outside Tham Meud cave, spiders built domed webs close to the ground, attached to the rocks or rock walls.Inside the cave, spiders were hanging from the apex of the dome while outside the cave they were sitting flat on the rock surface at the border of the web.No specimens were found in deeper parts of the cave (beyond about 20 m).

Justification of synonymy
The holotype of S. baso was compared directly with fresh specimens of S. miser originating from the type locality.The males were found to be identical in all relevant genital structures (cf.Figs 77-80).

Diagnosis
Easily distinguished from known congeners by morphology of male palp (strongly curved ventral trochanter apophysis; shapes of procursus and bulbal process Color.Carapace pale ochre; ocular area and clypeus dark brown; sternum black; legs light brown, with darker rings on femora (subdistally) and tibiae (proximally and subdistally); abdomen pale gray, with dark subcuticular marks lying above deeper white marks, with distinct ventral pattern consisting of three interconnected black marks.
Body.Habitus as in Fig. 74; ocular area elevated, each triad on short hump directed toward lateral; carapace with deep median furrow; clypeus with pair of rounded processes at rim, similar to S. kaeo sp.nov.but smaller (cf.Fig. 20); sternum wider than long (0.78 / 0.54), unmodified.Chelicerae as in Fig. 83, with pair of lateral processes, directed slightly toward posterior; without stridulatory ridges.
PalPs.As in Figs 81-82; coxa unmodified; trochanter with small retrolateral and ventral processes and distinctive long ventral apophysis, proximally attached to femur, distally strongly curved; procursus distally complex, with distinctive membranous and sclerotized elements; bulb with large proximal sclerite, with single complex process (Fig. 79) apparently containing sperm duct.legs.Without spines; with vertical hairs in higher than usual density on all tibiae; without curved hairs; retrolateral trichobothrium on tibia 1 at 9%; prolateral trichobothrium absent on tibia 1, present on other tibiae; tarsal pseudosegments indistinct, about 15 visible distally on tarsus 1.

Male (variation)
The holotype of S. baso is strongly bleached (cf.figs 127-128 in Huber 2005), but the distinctive male genital structures are indistinguishable from those of males from the type locality (Figs 77-80).In the fresh material from the type locality, all bulbs are rotated away from their natural position, which explains the differences between the palps of S. baso illustrated in Huber (2005) and the palps illustrated herein.

Female
In general similar to male; eye triads closer together (PME-PME distance: 150 µm); clypeus unmodified; leg tibiae with usual low density of vertical hairs.Tibia 1: 5.4 (missing or detached in other specimens).Epigynum slightly protruding (Fig. 35), with wide transversal sclerotized plate with pair of semicircular posterior extensions (Figs 34,38,84); internal genitalia as in Figs 39 and 85, pore plates contiguous (whether the sclerotized bars extending toward posterior are also provided with pores is not clear).

Distribution
Known from two localities in mainland Malaysia and Sumatra (Fig. 1); but see Addendum.