Review of the Australian and New Zealand orb-weaving spider genus Novakiella (Araneae, Araneidae)

The orb-weaving spider genus Novakiella Court & Forster, 1993 (family Araneidae Clerck, 1757) is reviewed to include two species, N. trituberculosa (Roewer, 1942) (type species, Australia and New Zealand) and N. boletus sp. nov. (Australia). Novakiella belongs to the informal, largely Australian ‘backobourkiine’ clade and shares with the other genera of the clade a single macroseta on the male pedipalp patella and a median apophysis of the male pedipalp that forms an arch over the radix. The proposed genus synapomorphies are the presence of a large basal conductor lobe expanding apically over the radix and the shape of the median apophysis, which extends into a basally directed, pointy projection. Males have an apico-prolateral spur on the tibia of the second leg that carries a distinct spine. Females have an epigyne with triangular base plate bearing transverse ridges and an elongate triangular scape, which is almost always broken off. The humeral humps of the abdomen are distinct. Novakiella trituberculosa build characteristic domeshaped webs; however, the foraging behaviour and web-shape of N. boletus sp. nov., currently only known from museum specimens,


Introduction
A recent multi-gene molecular phylogenetic analysis of the orb-weaving spider family Araneidae Clerck, 1757 recovered a well-supported Australasian clade informally termed 'backobourkiines' (Scharff et al. 2020). This clade includes species that would previously have been considered members of the Araneinae Clerck, 1757 (see Scharff and Coddington 1997), specifically those in the genera Acroaspis Karsch, 1878; Backobourkia Framenau, Dupér-ré, Blackledge & Vink, 2010;Carepalxis L. Koch, 1872;Novakiella Court &Forster, 1993 andPlebs Joseph &Framenau, 2012; but also Singa C. L. Koch, 1836, which has an almost global biogeographical distribution. In addition, the backobourkiines harbour Australian species apparently wrongly placed at the genus level, including some species currently listed in Araneus Clerck, 1757, Eriophora Simon, 1864 and Parawixia F. O. Pickard-Cambridge, 1904. All these genera were polyphyletic in Scharff et al.'s (2020) analysis, with Australian species placed in clades outside those that included the respective type-species or assumed close relatives from the same biogeographic region.
Morphologically, the backobourkiines are still poorly circumscribed, although a single macroseta on the male pedipalp patella and the median apophysis forming an arch over the radix seem to morphologically unite all genera against many clades previously considered part of the traditional Araneinae (Scharff and Coddington 1997;Scharff et al. 2020). An ongoing revision of Australian Araneidae suggests that the backobourkiines include many more described -generally misplaced at genus level -and undescribed species.
The orb-weaving spider genus Novakiella was not diagnosed in detail when initially described (sub Novakia Court & Forster, 1988). However, the original description of the genus pointed to a peculiar genital morphology and an unusual horizontal orb-web "drawn up into a cone by threads attached to the hub" built by the type species, N. trituberculosa (Roewer, 1942) (Court and Forster 1988, p. 123, figs 563-566). Therefore, Novakiella is easily distinguished from other backobourkiine spiders by their male pedipalp morphology with a distinct enlarged conductor lobe and the curved and basally projected median apophysis (Court and Forster 1988, figs 559, 560). Female epigynes have a subtriangular base plate with transverse ridges. The scape is wrinkled, elongate triangular and extends posteriorly past the base plate (Court and Forster 1988, figs 554, 555). The genus currently comprises the type species only, originally described from New Zealand but also known from Australia (Court and Forster 1988). The recent discovery of a second species from Australia allows a more comprehensive generic diagnosis of the genus in comparison to other backbourkiines. Therefore, we herein review Novakiella, provide a modern diagnosis against other backobourkiine orb-weavers and (re)describe its two species.

Materials and methods
Descriptions and terminology follow recent publications on Australian orb-weaving spiders (e.g., Framenau et al. 2010;Joseph and Framenau 2012;Castanheira et al. 2019). Colour patterns are described based on specimens preserved in 75% ethanol. Male pedipalps were expanded by alternatively submerging it for 10 min in 10% KOH and distilled water until fully expanded. Female genitalia of N. trituberculosa were dissected and cleared with lactic acid. Measurements are given in millimetres.
Images of specimens were taken in different focal planes with a Nikon D300 digital SLR camera attached via a C-mount adapter from LM-Scope (http://www. lmscope.com) to a Leica M16A stereomicroscope and combined with Auto Montage (vers. 5.02) software from Syncroscopy to increase depth of field. We used 2 Nikon R1C1 wireless speedlights instead of fibre optics to illuminate the exposures. The latter were used as guide-light for focusing. Microscopic images of cleared epigynes were taken in different focal planes (ca. 20-30 images) on a Leica DMC4500 digital camera mounted to a Leica M205C stereomicroscope and combined using the Leica Application Suite X, v. 3.6.0.20104. Images of expanded pedipalps of N. trituberculosa were taken with a BK Plus Laboratory System from Visionary Digital (Palmyra, PA, USA) equipped with a Canon EOS 7D camera. All photos were edited and mounted with Photoshop CC 2020.
Maps were compiled in the software package QGis v. 2.14.0 Girona (https://qgis.org/en/site/; accessed 21 January 2020). Geographic coordinates were extracted directly from original labels or the registration data as provided by the museums. When no detailed geographic information was available, localities were estimated based on Google Earth v. 9.1.39.3 (https://earth.google.com/web/; accessed 21 January 2021).
Diagnosis. The informal clade of the backobourkiines is well supported by the molecular phylogeny of Scharff et al. (2020), but the taxonomy and systematics of the species and genera within this clade are poorly resolved. Only three genera within the clade have been revised using modern taxonomic methods: Plebs, Backobourkia and Lariniophora Framenau, 2011. The genera Carepalxis and Acroaspis have not been revised and their putative synapomorphies remain unknown. It is therefore difficult to diagnose Novakiella against these genera. Other Australian backobourkiines included in Scharff et al. (2020) represent species that have clearly been misplaced in genera they do not belong to (i.e., Eriophora or Araneus) and these represent undescribed genera (in that study listed as "NGEN01" for Eriophora transmarina (Keyserling, 1865), "NGEN02" for Araneus recherchensis Main, 1954 and "NGEN05" for Araneus senicaudatus Simon, 1908). Until these species have been revised and placed in new or existing genera, Novakiella cannot be diagnosed against them.
Description. Medium-sized (TL males ca. 5-9, females 8-12) orb-weaving spiders with males on average slightly smaller than females. Carapace longer than wide, pear-shaped; cephalic area similar in shape in both sexes (Figs 1A, 3A, 4A, 6A); fovea longer than wide in males and wider than long in females, and with a dark spot in both sexes (Figs 1A, 3A, 4A, 6A); colouration (of ethanol preserved specimens) varying from reddish-brown to yellowish-brown, with black patches along carapace borders (Figs 1A, 3A, 4A, 6A). Eyes ringed in black, anterior median eyes largest, posterior eye row slightly recurved, lateral eyes almost touching, posterior lateral eyes separated from posterior median eyes by more than their diameter and located on small tubercles at the clypeus border (Figs 1A, 3A, 4A, 6A). Chelicera paturon with dark hue, fangs reddish-brown. Labium wider than long, subtriangular, with front end bulging and beige ( 4D-F, 5): male pedipalp patella with a single strong macroseta; paracymbium well-developed and hook-like; cymbium longer than wide; radix thick and elongated, reaching from the base of median apophysis to near the cymbium tip; conductor lobe conspicuous and projected apically, being composed of two distinct lobes (N. trituberculosa) or mushroom-shaped (N. boletus sp. nov.); terminal apophysis wider than long, rounded and tapering terminally; conductor well-developed, subquadrate; embolus uncapped, elongated, pointed and almost straight; median apophysis stout, with an acute basally pointing tip. Female genitalia (Figs 3C-E, 6C): epigyne plate wider than long, subtriangular; scape much longer than wide and extending posteriorly beyond plate (but length not known in N. boletus sp. nov.), generally broken off. Spermathecae spherical and occupying most of genital area.
Remarks. The study was conducted over many years and at different institutions and therefore imaging and descriptive work based on variable specimens (plural) availability at the time. This explains why the male N. trituberculosa is here redescribed based on two specimens; one imaged many years ago, but not measured, and the measurements added for a second specimen more recently.
Habitat preferences and life history. In Australia, mature males of N. trituberculosa were found between November and June, with peaks in January and April. Mature females were found all year round with the lowest numbers of records in November and December. Here, the species is mainly found in "pastoral habits" and constructs a horizontal orb-web amongst low grasses or weeds, with the centre pulled up by stabilizing threads. The webs are up to 0.1 m above ground. Additionally, habitat descriptions on specimen labels include "woodland", "open forest", "shrubs near ground", "in long grass", but the species also seems to occur in more disturbed habitats such as "among garden rubbish", "ex toilet", "walking on wall at night", "inside house on wall", "stationary on door knob". In New Zealand it is mostly found in pastoral habitats (Court and Forster 1988), which suggests that it is introduced.
Distribution. Novakiella trituberculosa has been recorded from all Australian states, except Northern Territory, south of ca. 22°S Latitude (Fig. 7). In New Zealand the species is more frequently found in the North Island but it has also been found in some South Island localities (Court and Forster 1988) (Fig. 8).
Diagnosis. Male N. boletus sp. nov. can be distinguished from N. trituberculosa by the weaker apico-prolateral spur on the tibia of leg II (Fig. 1E vs Fig. 5C) and the morphology of key pedipalp sclerites, specifically the mushroom-shaped conductor lobe (two-lobed in N. trituberculosa) (Fig. 1C vs Fig. 5F). Females of N. boletus sp. nov. differ from those of N. trituberculosa by details in the epigyne plate, specifically its transverse wrinkles that are more pronounced and mainly limited to the lateral margins in N. trituberculosa (Fig. 3C vs Fig. 6C).
Habitat preferences and life history. Mature males of N. boletus sp. nov. were collected between February and April, females were found in May and November. Habitat descriptions include Nothofagus cunninghamii (Myrtle Beech) forest, Eucalyptus amygdalina coastal forest, and "eucalypt forest with tree fern gully", suggesting this species occurs predominantly in temperate forests and rainforests.
Remarks. Males and female N. boletus sp. nov. have not been found together, but somatic features such as size range, carapace ( Distribution. This new species is only known from Australia, specifically New South Wales, South Australia, Victoria and Tasmania (Fig. 7).

Discussion
Araneidae are a highly diverse family with generally complex male pedipalp morphology which is traditionally used to infer phylogenetic relationships within this family as well as in other spiders (Scharff and Coddington 1997 and references therein). Developing homology hypotheses on the various pedipalp sclerites is a prerequisite to infer phylogenetic relationships using morphological data, subsequently tested by the phylogeny, but homology hypothesis based on the classical homology criteria of congruence, conjunction and similarity (Patterson 1988) is not straightforward. For example, the name paramedian apophysis has been used for structures in the male pedipalp that are very different in shape and with different position on the male pedipalp. For instance, the lobe of the conductor seen in Micrathena Sundevall, 1833 (Levi 1985, figs 6-9) was considered homologous with the separate sclerite seen in Gasteracantha Sundevall, 1833 (Levi, 1978, figs 83, 84) and named paramedian apophysis by Levi (1978Levi ( , 1985. The term paramedian apophysis was first used by Comstock (1910, pp. 179, figs 18, 19) for an extra sclerite in Eriophora ravilla (C.L. Koch, 1844). He writes "in this species there is an apophysis which like the median apophysis is joined by a flexible articulation to the tegulum within the cuplike cavity formed by the distal margin of the tegulum; this may be termed the paramedian apophysis". Scharff and Coddington (1997, fig. 95) tested the homology of the paramedian apophysis (in the broad definition of Levi) on a phylogeny based on morphological characters and found that the paramedian apophysis had developed several times independently within Araneidae. The same results are obtained if the character is mapped on the new molecular phylogeny of Scharff et al. (2020). Each presence of a paramedian apophysis therefore has to be considered individually and probably represent different non-homologous structures. Interestingly, a paramedian apophysis in the form of a separate sclerite inserting on the tegulum next to the median apophysis, is a possible synapomorphy for the clade called gasteracanthines in Scharff and Coddington (1997) and Scharff et al. (2020). In other backobourkiines (i.e. Backobourkia; Framenau et al., 2010, figs 6A, 10A;Plebs, Joseph & Framenau, figs 8A, 11A), the paramedian apophysis is clearly connected basally to the conductor and thus not homologous to the one in Micrathena, and could thus be better termed conductor lobe. Levi (1985) considered the paramedian apophysis as a synapomorphy that could group different genera like Eriophora, Parawixia, Alpaida O. Pickard-Cambridge, 1889 andWagneriana F. O. Pickard-Cambridge, 1904. In Novakiella, a basal conductor lobe is also present, but it is shaped very differently to that in other backobourkiines and the homology of a variety of structures in araneids in such a position remains unclear (Scharff and Coddington 1997). In Novakiella the conductor lobe is a very prominent structure that originates between the basis of the distal hematodocha and the stipes and fills the lateral space between the terminal apophysis and the embolus basis reaching far apically of the radix and connecting to the conductor from under the embolus. We initially thought it was a structure similar to the subterminal apophysis, which is a bubble-shaped structure that was first cited as a synapomorphy for Eustala Simon, 1895and Metazygia F. O. Pickard-Cambridge, 1904(Levi 1977 and then also cited for Larinia Simon, 1874 (Harrod et al. 1991), a member of the "Nuctenines" (sensu Scharff et al. 2020). However, it looks more related to the paramedian apophysis of Eriophora and Backobourkia cited above due to its origin at the base of the conductor and its shape. A large well-sclerotized transverse structure similar to the conductor lobe of Novakiella appears to be present in other "backobourkiine" genera, such as Acroaspis (see Framenau 2019: fig. 1B for A. lancearia (Keyserling, 1887). Testing homologies of the various pedipalp sclerites within the backobourkiines and to develop a generalized ground plan for this group will be a prerequisite to develop homology hypotheses to other major clades of the Araneidae as identified in Scharff et al. (2020). This can only be conducted once the apparently highly diverse backobourkiines have been taxonomically revised.
Novakiella trituberculosa was originally described from New Zealand, but Court and Forster (1988) considered the species to also occur in Australia, so the biogeographic origin of the genus remained ambiguous. The finding of a second species of Novakiella in Australia suggests that the genus evolved there and that N. trituberculosa is a natural or human-induced introduction to New Zealand. This is also consistent with Novakiella being part of the backobourkiines, a clade with likely Australian origin (Scharff et al. 2020). Likewise, Eriophora pustulosa (Walckenaer, 1841) is the only New Zealand species of a group of backobourkiines with a number of otherwise Australian representatives (VWF, PSC, CJV unpublished data).