1° primary

2° secondary

AC aciniform silk gland

AG aggregate silk gland

ALS anterior lateral spinneret

Col colulus

CY cylindrical (= tubuliform) silk gland

FL flagelliform silk gland

MaA major ampullate silk gland

MiA minor ampullate silk gland

MoPI modified piriform silk gland

PI piriform silk gland

PLS posterior lateral spinneret

PMS posterior median spinneret

T-A tartipore-accommodated

Located in the spinning fields of some spinnerets, tartipores form and function only during proecdysis (Table 1), acting within the underlying new exoskeleton as short conduits for those silk gland ducts that remain connected to spigots on the overlying old exoskeleton. An example of a tartipore that was preserved while functioning during proecdysis is shown in Fig. 4C.

Murphy and Roberts (2015) substituted the term ‘cicatrix’ for ‘tartipore’ on the grounds that “…tartipores…is misleading since they are not, in any sense, pores.” It is true that after ecdysis, after tartipores have served their purpose, they no longer function as pores, but during proecdysis they can be described as pores, if we accept ‘pore’ to be a minute opening in a surface. It would be disingenuous, however, not to acknowledge that the term ‘tartipore’ was initially applied, not to proecdysial tartipores functioning as pores, but to post-ecdysial, post-functional tartipores (Shear et al. 1989, Yu and Coddington 1990). Townley and Tillinghast (2003) continued this restricted, post-ecdysial use of the term and thus defined tartipores as cuticular scars, very much in accord with Murphy and Roberts’ (2015) use of ‘cicatrices’. The proecdysial, functional form of tartipores were referred to as ‘tartipore progenitors’. Put another way (with the objection of Murphy and Roberts in mind), these structures were recognized as definitive tartipores only once they no longer functioned as pores! In practice, maintaining a terminological distinction between proecdysial ‘tartipore progenitors’ and post-ecdysial ‘tartipores’ was found to be unnecessary, more confusing than clarifying, and, thus, ‘tartipore progenitors’ were subsequently also referred to as ‘tartipores’ (Townley and Tillinghast 2009). We concur that a single term should be applied to these conduits, whether functional or post-functional, and since they are not scars during proecdysis, nor pores after ecdysis, it can be argued that neither ‘cicatrix’ nor ‘tartipore’ is entirely fitting. But as ‘tartipore’ is already well-established, we will continue to use it.

Only the ducts of some silk glands remain connected to the old exoskeleton during proecdysis and result in tartipores forming in the developing exoskeleton around the ducts. These silk glands are said to be tartipore-accommodated (T-A) and can have silk drawn from them during proecdysis. The ducts of non-tartipore-accommodated (non-T-A) silk glands lose their connection to the old exoskeleton around apolysis, do not give rise to tartipores, and therefore the spider cannot draw silk from them during proecdysis. In some spider clades, such as the ecribellate haplogyne Synspermiata (Platnick et al. 1991, Michalik and Ramírez 2014, Garrison et al. 2016, Wheeler et al. 2016), all silk glands are non-T-A. The two Australomimetus species examined in this study have both T-A and non-T-A versions of major ampullate, minor ampullate, and piriform silk glands (MaA, MiA, PI). A. spinosus also has T-A and non-T-A aciniform silk glands (AC), whereas A. djuka only has non-T-AAC, and the female-specific cylindrical silk glands (CY) are exclusively non-T-A in both species. Those MaA and MiA that are non-T-A are called primary (1°) MaA/MiA while those that are T-A are called secondary (2°) MaA/MiA. This nomenclature reflects both the larger size of the 1° MaA/MiA and their use over a greater portion of juvenile stadia: from each ecdysis to the next apolysis, as opposed to the 2° MaA/MiA that appear to function, at least in Araneus, only very shortly before and during proecdysis (Townley et al. 1993, Tillinghast and Townley 1994). Typical of araneoids, in Australomimetus, 1° MaA/MiA function in juveniles and adults while 2° MaA/MiA function only in juveniles (and some Australomimetus species, like A. djuka, lack functional 2° MiA entirely).

Depending on species, sex, spinneret, instar, and maturity, certain spinning fields consistently contain vestigial spigots, termed nubbins, at sites occupied by functional spigots either in earlier instars (ontogenetic nubbins; often present in adults, such as 2° MaA/MiA nubbins) or in ancestors (phylogenetic nubbins). Examples of both these types of nubbins were observed during this study and are noted below. In the literature up through about 1994 (and sometimes beyond), the structures now identified as 2° MaA/MiA tartipores were generally referred to as nubbins [non-vestigial-type nubbins in Townley et al. (1993)] because the terms ‘nubbin’ and ‘tartipore’ were originally applied to non-spigot protuberances in spinning fields on the basis of whether or not they can be individually specified (Coddington 1989, Yu and Coddington 1990). Later, once the very different developmental origins of nubbins and tartipores were better understood, ‘nubbin’ was restricted to its current meaning of vestigial spigot (Townley 1993, Townley and Tillinghast 2003). The above terms are summarized in Table 1.

With our focus on external features of the exoskeleton, we consider ecdysis to be the event that separates instars (the spiders themselves) and stadia (the periods of time between ecdyses or from ecdysis to death) since, from an external view, it is at ecdysis that there is a change of exoskeleton. Following apolysis, we apply the term pharate instar to only that portion of a proecdysial spider from the new exoskeleton inward. Thus, using the example of a specimen presented extensively in this paper, a 4^th instar that undergoes apolysis, considered as a whole, remains a 4^th instar throughout proecdysis, and the pharate 5^th instar constitutes only a part (albeit the predominant part) of this proecdysial 4^th instar. As proposed by Downes (1987), and adopted by others (e.g., Townley and Tillinghast 2009, Wolff and Hilbrant 2011, Hilbrant et al. 2012, Mittmann and Wolff 2012), when an embryo hatches, the postembryo is released, and ecdysis in the postembryo, including the discarding of its exuvium (shed exoskeleton), produces the 1^st instar and marks the start of the 1^st stadium. The molt between the postembryonic stage and 1^st stadium is taken to be the 1^st molt (Downes 1987). Subsequent ecdyses produce consecutively numbered instars and stadia. Fig. 1 schematically summarizes these terms as used in this paper. Though some definitions of instar/stadium exclude adults (Carlson 1983), we distinguish spiders that molted 5 times before becoming adults from those that molted 6 times by referring to 5^th stadium and 6^th stadium adults, respectively.

Figure 1.

Developmental terms as applied in this paper, largely based on the proposal of Downes (1987). Indicated relative durations of stages are not intended to be accurate. Apo, apolysis; Ecd, ecdysis; Pro, proecdysis.

Specimens of Australomimetus spinosus were reared in captivity from egg sacs collected at Yandin-Lookout, Western Australia (30°46’23”S, 115°36’31”E) in August 2014. Australomimetus spinosus is the only mimetid spider that occurs in the Yandin Nature Reserve which comprises open eucalypt woodland and Xanthorrhoea shrubland on rocky outcrops. All egg sacs, containing 10-25 eggs each, were collected from under large rocks and boulders together with females. In the lab, up to four egg sacs were placed in glass jars capped with nylon and the emerging spiderlings were fed with prey spiders (Theridiidae) collected in the field. To avoid cannibalism in A. spinosus post-molt, the spiderlings were progressively separated from the 3^rd stadium onwards and 3-5 specimens of later instars were reared in small jars with sufficient prey and some paper tissue. Specimens of all life stages were euthanized in 100% ethanol between August 2014 and January 2015, then stored in a freezer for further analyses. Exuvia of representative specimens were also preserved in ethanol post-molt. Additional museum specimens of A. spinosus were also available from Western Australia, Queensland, and New South Wales, collected between January 1992 and July 2014 and stored in 70-75% ethanol. All specimens of A. djuka were from the collections of the Western Australian Museum (WAM), and stored in 70-75% ethanol. This species has a larger body size but also considerably higher moisture requirements than A. spinosus. It is endemic to the southwest of Western Australia and has been collected from limestone caves in Yanchep National Park but also dense bark accumulations in undisturbed karri (Eucalyptus diversicolor) forest. Additional information on material examined is given in Suppl. material 1.

A total of 11 sets of spinnerets (6 spinnerets/set; 1 set/abdominal exoskeleton) from A. djuka and 63 sets of spinnerets from A. spinosus were examined by SEM (Table 2). These included one or more sets from each instar that has a functioning spinning apparatus (1^st to 6^th instars). Of the 63 A. spinosus sets, 49 were from 1^st to 5^th instars, all juveniles from Western Australia and laboratory-raised. The other 14 sets, which included those from all 9 A. spinosus adults that were examined, were from spiders field-collected either in Western Australia or in Queensland or New South Wales (Table 3). The 49 sets of spinnerets from laboratory-raised specimens included 7 from exuvia of 2^nd instars and 7 from exuvia of 3^rd instars. Later spinnerets from at least some of the spiders that shed these exuvia were also examined. In addition, one laboratory-raised 4^th instar in proecdysis yielded two sets of spinnerets, one from the old exoskeleton, one from the new (pharate 5^th instar’s) exoskeleton. Thus, an unknown number of spiders, but probably ≤ 37, gave rise to the 49 sets of spinnerets from laboratory-raised spiders. With the exception of the two sets from the proecdysial 4^th instar, we do not know which sets of spinnerets came from the same individual. Overall, this is one of the more comprehensive spinneret data sets on juvenile spiders to date, covering all life stadia, 1^st to adult, in these two species.

Intact sets of spinnerets, including underlying tissues and small amounts of surrounding exoskeleton, were cut off ethanol-preserved specimens of the two Australomimetus species using Vannas spring scissors (Fine Science Tools, 15000-08) while being viewed through an Olympus SZX12 stereo microscope. They were immersed in Novex® Tris-glycine sodium-dodecyl-sulfate (SDS) buffer (ThermoFisher Scientific, LC2675) at 2X-strength for at least 3 days at 4°C. If underlying tissues did not show signs of partial solubilization in the buffer (expansion, increased translucence, separation from exoskeleton), incubation was continued at room temperature for up to 4 months. After a brief (5 min) rinse in commercial contact lens saline solution, spinnerets were further cleaned, and remaining underlying tissues digested, in Oxysept® Disinfecting Solution containing one Oxysept® Neutralizing Tablet and one Ultrazyme® Enzymatic Cleaner Tablet (all Abbott Medical Optics, intended for use with soft contact lenses) according to manufacturer’s instructions. This treatment exposed spinnerets to 3% H₂O₂ and subtilisin A and was allowed to proceed until underlying tissues were no longer visible, occasionally requiring overnight incubation. Spinnerets were then transferred to a Petri dish containing a wax layer overlaid with the above Novex® buffer. The tips of pins of various sizes projected 1-3 mm upward from the wax layer (Townley and Tillinghast 2009). Each spinneret in a set was pulled down onto a pin tip of appropriate diameter to expand the spinneret as fully as possible. As Fig. 9G demonstrates, expansion was not always entirely successful. Spinnerets were then dehydrated through an ethanol series (30%, 50%, 70%, 85%, 95%, 100%, 100%; at least 1 day in each), followed by transfer into microporous specimen capsules [11 x 12 mm, 78 µm pore size, 1 set of spinnerets/capsule; Electron Microscopy Sciences (EMS), 70187-10] that were submerged in fresh 100% ethanol. All transfers of spinnerets from one solution to another were accomplished using transfer pipets so that severed spinnerets were always in liquid from their initial immersion in Novex® buffer to their transfer to the microporous capsules. Capsules were placed in the ethanol-filled chamber of a Tousimis Samdri-795 critical point dryer and dried. Spinnerets were then affixed to aluminum SEM specimen pin mounts (EMS, 75160) using double-sided conductive carbon adhesive tabs (EMS, 77825-09) and PELCO® conductive carbon glue (Ted Pella Inc., 16050), sputter-coated with about 20 nm gold/palladium, and examined on a Tescan Lyra3 XMU field-emission SEM (6 kV, beam intensity 6).

Some 2^nd and 3^rd stadia spinnerets were obtained from exuvia that were initially stored in 70-75% ethanol, but were subsequently immersed in the aforementioned SDS-containing Novex® buffer for at least 2 months at 4°C. These were prepared like the above spinnerets from intact specimens except that they were not treated with the H₂O₂/subtilisin A solution and they were not intentionally separated from the rest of the exuvium, though separation often resulted nonetheless from the act of pulling the spinnerets down onto the tips of pins to expand them.

Diameters of AC spigot openings, presented in Figs 7E, G, 8E, 9J, K, and 10F, were obtained using the measurement tool in Tescan’s Lyra 3 Control Software. Because borders of some openings deviated noticeably from a circle, each value given in a figure is the mean of two measurements: the opening’s widest diameter and that perpendicular to the widest.

Copulatory structures, or evidence of their development, were the primary features consulted in determining the sex of adults and penultimate instars. The sex of younger specimens, as early as 3^rd instars of A. spinosus and 4^th instars of A. djuka, was primarily determined by the presence (female) or absence (male) of CY spigots on PMS and PLS. As detailed in the Results (‘Cylindrical silk gland…’), the full complement of CY spigots was not present in the majority of 3^rd stadium female A. spinosus, raising the possibility that some 3^rd stadium females do not exhibit any CY spigots and leaving some uncertainty in our assignment of sex for a few 3^rd stadium individuals. CY spigots appear to be absent in female as well as male 1^st and 2^nd instars. Thus, sex was not determined for these instars.

Convention 1. Each of the three pairs of spinnerets (ALS, PMS, PLS), split midsagittally, form right/left mirror-images (in overall arrangement) of each other. Some spinnerets depicted in the figures were from the right side of the opisthosoma, some from the left. But to facilitate comparisons, those from the right side have been flipped in Microsoft® PowerPoint® 2016 so that they also appear to be left spinnerets. Figure legends, however, state their true handedness (right, left). Figs 3–11 are divided into three groups by spinneret pair; images in Figs 3–5 show ALS, Figs 6–8 show PMS, and Figs 9–11 show PLS. Within each group, the first figure depicts the spinneret in 1^st to 3^rd instars of both A. djuka and A. spinosus, while the second and third figures show the same spinneret type in later instars of just A. spinosus and just A. djuka, respectively.

Convention 2. On each anterior lateral spinneret (ALS) and posterior lateral spinneret (PLS), spinning fields are restricted to the more distal of the two segments composing the spinneret. Posterior median spinnerets (PMS) are single-segmented with spigots apically placed. Fig. 2A provides an overview of a set of spinnerets (though the ALS and PLS from one side of the opisthosoma are only partially visible), with boxes approximating the regions and perspectives shown in Figs 3–11. For ALS and PLS, essentially only the medial face of the distal segment is shown in Figs 3–5 and Figs 2B, 9–11, respectively. For PMS, all or nearly all of the spinneret is shown in Figs 6–8, most often from a perspective roughly perpendicular to the opisthosoma surface. Note, however, that the boxed PMS in Fig. 2A depicts a right PMS whereas Figs 6–8 depict left PMS (even though some are actually right PMS, see Convention 1).

Figure 2.

Spinnerets of Australomimetus spinosus. A. Overview of spinnerets. Boxes indicate approximate regions and perspectives shown in many of the images in Figs 3–11. Upper spinnerets from right side of opisthosoma (image flipped). B. Division of PLS into three AC regions: non-T-AAC (AC spigot column between two white curves), anterior T-AAC (AC spigots and tartipores left of left white curve), and posterior T-AAC (AC spigots and tartipores right of right white curve). AC tartipores indicated by unlabeled arrows. White curves in Figs 9, 10 likewise separate these three regions. Left PLS; all spigots are those of AC. Scale bars: A 100 µm; B 20 µm. See also Abbreviations and terminology (‘Spinning apparatus abbreviations’) and Results (Convention 3 in ‘Conventions applied…’).

Figure 3.

ALS of 1^st to 3^rd instars of Australomimetus spinosus and Australomimetus djuka. A–C. A. spinosus. D–F. A. djuka. A, D. Boxed regions, magnified in insets, show putative 2° MaA tartipore primordia. Unlabeled arrows indicate the four PI spigots present in 1^st instars (two T-A, two non-T-A); note increased PI spigot numbers in later instars. B, E. Unlabeled arrows indicate the two PI tartipores present in 2^nd instars; note increased PI tartipore numbers in later instars. A, C, E, F. Left ALS. B, D. Right ALS (image flipped). 1°, 1° MaA spigot; 2°, 2° MaA spigot; T, 2° MaA tartipore. All scale bars 20 µm. See also Abbreviations and terminology (‘Spinning apparatus abbreviations’) and Results (Convention 3 in ‘Conventions applied…’).

Convention 3. Figures include a label indicating the instar from which the spinneret shown was obtained (1^st to 6^th; Ph, pharate instar) and, if known, the spider’s sex (♀, ♂) followed by a state of maturity subscript (J, juvenile; A, adult). For example, 4^th ♀_J indicates a juvenile female 4^th instar. Sex was unknown only for 1^st and 2^nd instars, though uncertain for some 3^rd instars [see Results (‘Cylindrical silk gland…’)].

Convention 4. One of the new findings of the present study is that A. spinosus, unlike A. djuka and five other Australomimetus species previously examined (Harms and Harvey 2009b, Townley and Tillinghast 2009), has T-AAC in addition to non-T-AAC emptying on the PLS. The T-AAC spigots and tartipores occur both anterior and posterior to a more centrally located column of non-T-AAC spigots on each PLS, as shown in the example in Fig. 2B. Thus, we make a distinction between anterior and posterior T-AAC spigots and tartipores in Table 3, Figs 2B, 9, 10, and the text.

Convention 5. As noted in ‘Spinnerets examined’ in Materials and methods, in only one instance were two consecutive sets of spinnerets from the same individual identified; 4^th and pharate 5^th instar sets from a proecdysial 4^th stadium juvenile male A. spinosus, separated from one another during enzyme cleaning and examined separately. All of the colorized micrographs (Figs 4A, B, 7A, B, 10A–D) are from this male’s two sets of spinnerets. The color coding is as follows:

Figure 4.

ALS of 4^th to 6^th instars of Australomimetus spinosus. A-C.ALS from one juvenile specimen, well into proecdysis. Color coding explained in Results (Convention 5 in ‘Conventions applied…’). A. Old exoskeleton (all spigots but one damaged; 1° MaA spigot torn out entirely, but its former location is colored red). B. New exoskeleton directly beneath (A). C. New exoskeleton, opposite ALS from (A, B), demonstrating utility of a tartipore. Unlabeled arrow to 2° MaA duct (severed at distal end) still accommodated by 2° MaA tartipore (T). Before old and new exoskeletons were separated, this duct emptied on a 2° MaA spigot on old exoskeleton (like that in (A) but on opposite ALS). D-G. Adults. Presumed pair of non-T-API spigots (stars) on each ALS in males (D, E) were morphologically similar to T-API spigots (unlabeled PI spigots) in both sexes and to non-T-API spigots (stars) in females (F, G): no MoPI spigots. C. Left ALS. A, B, D-G. Right ALS (image flipped). 1°, 1° MaA spigot; 2°, 2° MaA spigot; N, 2° MaA nubbin (ontogenetic); T, 2° MaA tartipore. Scale bars: A-D, F 20 µm; E, G 5 µm. See also Abbreviations and terminology (‘Spinning apparatus abbreviations’) and Results (Convention 3 in ‘Conventions applied…’).

Red: Spigots of non-T-A silk glands that were functional during the 4^th stadium and would have been functional again during the 5^th and 6^th (adult) stadia, though not during proecdyses. Thus, they were not functional at the time the spider was preserved.

Purple (Fuchsia): Spigots of non-T-A silk glands that would have been functional for the first time during the 5^th stadium, though not during proecdysis, and then again during the 6^th (adult) stadium. Only certain AC spigots (Fig. 10C, D) fall into this category.

Blue: Spigots and tartipores of T-A silk glands that were functional during the 4^th stadium, at least during proecdysis (and presumably also well prior to apolysis for at least some of these silk glands, such as PI). The PI and AC members of this group would have been functional again during the 6^th (adult) stadium, while the 2° MaA and 2° MiA members would have atrophied, their spigots replaced by nubbins. In the 5^th instar, the presence of these silk glands would have been indicated externally only by post-functional tartipores.

Yellow: Spigots and tartipores of T-A silk glands that would have been functional during the 5^th stadium and had previously been functional during the 3^rd stadium, at least during proecdyses (and presumably also well prior to apolysis for at least some of these silk glands, such as PI). None of these would have been functional during the 6^th (adult) stadium. In the 4^th instar, the presence of these silk glands was indicated externally only by post-functional tartipores.

Green: Spigots of T-A silk glands that would have been functional for the first time during the 5^th stadium, at least during proecdysis (and presumably also well prior to apolysis for at least some of these silk glands, such as PI). None of these would have been functional during the 6^th (adult) stadium.

Convention 6. Tables 2 and 3 compile spigot, tartipore, and nubbin data from all examined spinnerets that yielded reliable counts. Data for silk gland types with multiple spigots and tartipores per spinneret (at least in some stadia)―namely, PI data from both species (Table 2) and PLSAC data from A. spinosus (Tables 2, 3)―are color coded to guide the eye in making comparisons between spigot counts from one stadium and tartipore counts from the next. For example, in Table 2 it can be seen that in female A. spinosus, 3^rd instars had a mean of 12.3 PI spigots/ALS and, shaded in the same color, 4^th instars had 10.3 PI tartipores/ALS, consistent with two PI spigots/ALS in 3^rd instars being outlets for non-T-API, the remainder outlets for T-API. PLSAC spigot and tartipore data from A. spinosus are presented first on a ‘counts per PLS’ basis (Table 2) and then divided on a ‘counts per PLS region’ basis (Table 3), where three regions are recognized on each PLS: an anterior T-AAC region, a central non-T-AAC region, and a posterior T-AAC region (Fig. 2B). The data in Table 3 are further segregated by geography [eastern Australia (Queensland, New South Wales)/Western Australia].

Preserved spiders and exuvia were assigned to specific instars/stadia based on numbers of spigots and tartipores from the three pairs of spinnerets and, for A. spinosus, the position of the 2° MiA tartipore (Table 2). Though we believe at least most of these assignments are accurate, a small amount of uncertainty remains for later instars (beyond the 3^rd), especially in A. djuka. This is because 1) there is variation in the number of molts completed before maturity is attained, 2) ranges of spigot/tartipore counts can overlap, especially between later instars, 3) we have considerably more data for A. spinosus than A. djuka, and 4) spinnerets of A. spinosus are more informative than those of A. djuka thanks to the presence of 2° MiA and T-APLSAC in the former, but not the latter.

Based on examination of spinnerets from nine adults, both female and male A. spinosus have the potential to become adults after either five or six molts [Table 2; see Abbreviations and terminology (‘Terminology’) and Fig. 1 for numbering of molts, instars, stadia]. Consistent with this conclusion, palps swollen to different extents indicated that one of eight male 4^th instars was in the penultimate stadium, the rest in the antepenultimate stadium, and all four juvenile male 5^th instars were in the penultimate stadium. With data from only three A. djuka adults, we can only say that at least some males become adults after five molts and at least some females become adults after six molts. Consistent with this finding, all three juvenile 5^th instars examined were females (Table 2). For both species, but especially A. djuka, examinations of additional specimens could very well reveal greater variability with respect to the number of molts to maturity for one or both sexes. Overall, there seems to be some variability regarding the number of molts to maturity, with food availability and space being potential factors determining the number of molts.

Contrary to what was observed on the other two pairs of spinnerets, described below, there were no substantial differences between ALS of A. spinosus and A. djuka. Both had single 1° and 2° MaA spigots on each ALS in all instars, with the exception that the 2° MaA spigot was replaced by a 2° MaA nubbin in adults (Table 2, Figs 3–5). Such nubbins were thus ontogenetic nubbins. Both species had four PI spigots/ALS in all examined 1^st instars (Table 2, Fig. 3A, D) that were outlets for two T-API and two non-T-API, as demonstrated by two PI tartipores/ALS in all 2^nd instars (Table 2, Fig. 3B, E). With each molt, numbers of PI spigots and tartipores increased (Table 2, Figs 3–5), but the number of non-T-API very likely remained constant (two/ALS) through all stadia. The best evidence for this came from a proecdysial 4^th stadium male A. spinosus [see Results (Convention 5 in ‘Conventions applied…’)] that allowed us to compare the right ALS on both old (Fig. 4A) and new (Fig. 4B) exoskeletons (the left ALS of the old exoskeleton was too damaged for such a comparison). Of 16 PI spigots on the 4^th instar’s old exoskeleton (Fig. 4A, blue and red PI spigots), 14 (blue) could be paired with 14 PI tartipores in the pharate 5^th instar’s exoskeleton (Fig. 4B, blue PI tartipores), leaving two non-T-API spigots (Fig. 4A, red PI spigots) that, though non-functional during proecdysis, were functional earlier in the 4^th stadium and would have been functional after ecdysis during the 5^th stadium (Fig. 4B, red PI spigots).

Figure 5.

ALS of 4^th to 6^th instars of Australomimetus djuka. A. Juvenile. B–E. Adults. Presumed pair of non-T-API spigots (stars) on each ALS in males (B, C) were morphologically similar to T-API spigots (unlabeled PI spigots) in both sexes and to non-T-API spigots (stars) in females (D, E): no MoPI spigots. A, D, E. Left ALS. B, C. Right ALS (image flipped). 1°, 1° MaA spigot; 2°, 2° MaA spigot; N, 2° MaA nubbin (ontogenetic); T, 2° MaA tartipore. Scale bars: A, B, D 20 µm; C, E 10 µm. See also Abbreviations and terminology (‘Spinning apparatus abbreviations’) and Results (Convention 3 in ‘Conventions applied…’).

When PI spigot numbers in specimens at one stadium were compared with PI tartipore numbers in specimens at the next stadium, there were further indications of a constant two non-T-API/ALS in all instars. For example, in A. spinosus, PI spigots in 13 2^nd instars and PI tartipores in 15 3^rd instars numbered from 7-10 (Ῡ = 8.2)/ALS and 5-8 (Ῡ = 6.3)/ALS, respectively (Table 2). Another such comparison between 3^rd and 4^th instars of A. spinosus was noted earlier [Results (Convention 6 in ‘Conventions applied…’)]. Not all of these comparisons suggested constancy in the number of non-T-API across stadia, but this can probably be attributed to variation in the number of T-API at each stadium and the low number of observations involved in some comparisons.

In the proecdysial 4^th stadium male A. spinosus, nine PI tartipores were present in the old exoskeleton (Fig. 4A, yellow PI tartipores), indicating the positions of T-API spigots in the exoskeleton of the 3^rd instar. Interestingly, a group of nine PI spigots in the pharate 5^th instar’s exoskeleton (Fig. 4B, yellow PI spigots) were arranged, relative to one another and to the two non-T-API spigots (red), very much like the PI tartipores in the 4^th instar. This is compelling evidence that the same T-API that functioned during the 3^rd stadium were about to return to service during the 5^th stadium, and, by extension, that two of these nine T-API had functioned for the first time during the 1^st stadium. In keeping with the number of T-API spigots increasing after each molt (Table 2, Figs 3–5), the pharate 5^th instar’s ALS contained, in addition to spigots of the nine returning T-API (Fig. 4B, yellow), 13 spigots of T-API (Fig. 4B, green) that were about to function, apparently for the first time, during the 5^th stadium. These were located at the periphery of the PI spinning field.

In Fig. 4A, B, the spigot of the 1° MaA has been colored red to emphasize that this silk gland was non-T-A, like the two red-colored non-T-API associated with each ALS; all three apparently functional during every stadium and at all times except during proecdysis. Likewise, spigots and tartipores of 2° MaA have been colored blue or yellow to emphasize that they were T-A, like the blue- or yellow-colored T-API, and were functional at least during proecdysis. Any single T-A silk gland, however, functioned only during alternate stadia, i.e., either during even-numbered stadia (Fig. 4A, blue), with an accommodating tartipore forming in odd-numbered pharate instars (Fig. 4B, blue), or during odd-numbered stadia (Fig. 4B, yellow), with an accommodating tartipore forming in even-numbered pharate instars (which, in the case of the silk glands indicated in yellow in Fig. 4B, would have been the pharate 6^th instar, just as the yellow tartipores in Fig. 4A had earlier formed in the pharate 4^th instar to accommodate silk glands functioning during proecdysis in the 3^rd stadium). Two colors (blue, yellow) thus signify different groups of silk glands functioning in alternation with one another.

Postembryos were not examined, but all indications were that functional spigots first appeared in 1^st instars. Thus, in both species, PI and 2° MaA tartipores first appeared in the exoskeletons of 2^nd instars (Fig. 3B, E). MaA spinning fields of 1^st instars contained a structure we interpret to be a 2° MaA tartipore primordium (Fig. 3A, D). These occupied positions similar to those taken by 2° MaA tartipores in the exoskeletons of 2^nd (Fig. 3B, E) and later instars (Figs 3C, F, 4, 5), and were only observed in the exoskeletons of 1^st instars. Though generally round and containing a depression or pore, the precise appearance of these presumed primordia varied within each species and within individuals. We do not attach significance to the difference in morphology exhibited in the examples shown in Fig. 3 A, D.

In some mimetid males, maturity brings the appearance of a pair of modified PI (MoPI) spigots to each ALS that differ morphologically from the other PI (Townley and Tillinghast 2009, Townley et al. 2013, Benavides and Hormiga 2016). It has been hypothesized that the two pairs of non-T-API, functional from the 1^st stadium, transform in the pharate final instar into the MoPI, such that their spigots at least, and presumably their corpora as well, become recognizably different from those of other PI (Townley and Tillinghast 2009, Townley et al. 2013). In none of the adult males examined in this study (five A. spinosus, two A. djuka) were MoPI spigots observed. Representative ALS from adult male A. spinosus and A. djuka are shown in Figs 4D, E, 5B, C, respectively, with spigots in the positions of non-T-API spigots indicated by stars. These showed no obvious differences to surrounding T-API spigots, nor to non-T-API spigots in adult females of these species (Figs 4F, G, 5D, E, respectively).

PMS of the two species differed from the 1^st stadium. In A. djuka, 1^st and 2^nd instars had one AC spigot/PMS, increasing to two in a 3^rd instar, while A. spinosus invariably started off with two AC spigots/PMS in 1^st instars (Table 2, Fig. 6A, E–H). These and all PMSAC in both species were non-T-A. More significant were differences in MiA complements. Both species had a single 1° MiA spigot on each PMS from the 1^st to the final stadium, but only A. spinosus also had one 2° MiA spigot/PMS at all juvenile stadia (Table 2, Figs 6–8). Observations on the 2° MiA in A. spinosus matched those of the 2° MaA: 1) the 2° MiA spigot was replaced by an ontogenetic 2° MiA nubbin in adults (Fig. 7C, D, F), 2) a 2° MiA tartipore first appeared in 2^nd instars (Fig. 6B), and 3) a structure we interpret to be a 2° MiA tartipore primordium was adjacent to the 2° MiA spigot on the PMS of all 1^st instars (Fig. 6A), but no later instars (Figs 6B–D, 7). Nor was such a structure seen on PMS of 1^st instars of A. djuka (Fig. 6E, F).

Figure 6.

PMS of 1^st to 3^rd instars of Australomimetus spinosus and Australomimetus djuka. All unlabeled arrows point to non-T-AAC spigots. A–D. A. spinosus. A. Boxed region, magnified in inset, shows putative 2° MiA tartipore primordium. B–D. 2° MiA spigot (2°) and 2° MiA tartipore (T) switched positions from 2^nd to 3^rd stadium. This switching continued in later stadia (Fig. 7). C.CY spigots first appeared in 3^rd instars of A. spinosus. Same CY spigot (C) and closest AC spigot (unlabeled arrow) shown from two perspectives. E–H. A. djuka. E, F. Same PMS shown from two perspectives. Note phylogenetic 2° MiA nubbin (N) in position occupied by 2° MiA spigot in A. spinosus (A). B, C, E, F. Left PMS. A, D, G, H. Right PMS (image flipped). 1°, 1° MiA spigot; 2°, 2° MiA spigot; C, CY spigot; N, 2° MiA nubbin (phylogenetic); T, 2° MiA tartipore. All scale bars 20 µm. See also Abbreviations and terminology (‘Spinning apparatus abbreviations’) and Results (Convention 3 in ‘Conventions applied…’).

The absence of functional 2° MiA in all examined specimens of A. djuka appears to represent a secondary loss: an apparent phylogenetic 2° MiA nubbin was observed on all four examined PMS from two 1^st instars of A. djuka (Table 2, Fig. 6E, F). This nubbin occupied the same position taken by the 2° MiA spigot in 1^st instars of A. spinosus (cf. Fig. 6A, E) and was not observed in later instars (Figs 6G, H, 8A–D).

On both old and new exoskeletons from the proecdysial 4^th stadium male A. spinosus [see Results (Convention 5 in ‘Conventions applied…’)], spigots of non-T-A silk glands on each PMS included all (two) AC spigots and the 1° MiA spigot (Fig. 7A, B, red). These silk glands were nonfunctional during this and earlier proecdyses, but were potentially functional between ecdysis and apolysis during this and all previous stadia and would have been functional again during the 5^th and 6^th (adult) stadia. The only T-A silk glands were two 2° MiA associated with each PMS, one functional during proecdyses of the 2^nd and 4^th stadia (Fig. 7A, blue), the other functional during proecdyses of the 1^st, 3^rd, and (upcoming) 5^th stadia (Fig. 7B, yellow). This arrangement again matched that of ALS-associated 2° MaA. The blue 2° MiA tartipore in Fig. 7B was accommodating the duct of the 2° MiAfunctioning during this 4^th stadium proecdysis, while the yellow 2° MiA tartipore in Fig. 7A had accommodated the other 2° MiA during proecdysis in the 3^rd stadium.

Figure 7.

PMS of 4^th to 6^th instars of Australomimetus spinosus. All unlabeled arrows point to non-T-AAC spigots. A, B.PMS from one juvenile specimen, well into proecdysis. Color coding explained in Results (Convention 5 in ‘Conventions applied…’). A. Old exoskeleton. B. New exoskeleton directly beneath (A). C. Adult female. With one exception [see Results (‘PMS-Sexual and geographic…’) and Discussion (‘AC spigot numbers…’)], adult females had two AC spigots/PMS. D-G. Adult males from eastern Australia (New South Wales) (D, E) and Western Australia (F, G). AC spigots magnified in (E, G) to show wider openings on ‘late’ AC spigots (L) relative to neighboring ‘pioneer’ AC spigots (P) [see Results (‘PMS-Sexual and geographic…’)]; measured diameters of openings given in nm [see Materials and methods (‘SEM of spinnerets…’)]. In (G), distal ends of AC spigot shafts magnified in insets (connected by dashed lines), with each spigot’s counterpart on the opposite PMS shown in an adjacent inset. H–J. Ontogenetic changes in CY spigot (C) morphology; lateral view. Note changes in width relative to nearest AC spigot (arrow). A–G, I, J. Left PMS. H. Right PMS (image flipped). 1°, 1° MiA spigot; 2°, 2° MiA spigot; C, CY spigot; L, ‘late’ AC spigot; N, 2° MiA nubbin (ontogenetic); P, ‘pioneer’ AC spigot; T, 2° MiA tartipore. Scale bars: A-D, F 20 µm; E, G-J 10 µm. See also Abbreviations and terminology (‘Spinning apparatus abbreviations’) and Results (Convention 3 in ‘Conventions applied…’).

AC associated with the PMS appeared to present an example of sexual dimorphism in A. spinosus, though examination of additional specimens will be needed to establish the consistency of certain details. With one exception, the number of AC spigots in females was a constant two/PMS at all stadia (Table 3, Figs 6C, 7C). Adult males and one 4^th (antepenultimate) stadium male, in contrast, had between 3-5 AC spigots/PMS (Table 3). The additional AC spigots in these males (hereafter ‘late AC spigots’) were anterior to the two AC spigots present from the 1^st stadium (hereafter ‘pioneer AC spigots’) (Fig. 7D–G), close to the position of the CY spigot in females (Fig. 7C). Late AC spigot morphology did not differ greatly from that of pioneer AC spigots, though in the single examined adult male from Western Australia (with three AC spigots on each PMS, Table 3), the distal end of the base of the late AC spigot was narrower and with a less pronounced rim than those of the pioneer AC spigots (Fig. 7G). One consistent morphological difference between late and pioneer AC spigots was a wider opening on the late AC spigots (Fig. 7E, G). This applied to all six males that had at least three AC spigots/PMS (which included all five examined adult males), one from Western Australia, five from eastern Australia (Table 3). The single exceptional female was a 5^th stadium adult that had three AC spigots on one PMS (two on the other PMS), with the late AC spigot separated from the pioneer AC spigots by a gap (about the diameter of an AC spigot) not seen in any male. As in males, the opening of the late AC spigot was wider than those of the pioneer AC spigots.

AC spigot data from A. djuka did not suggest a clear sexual dimorphism like that indicated in A. spinosus, though there were some similarities. Females, including the single adult female examined, usually had two AC spigots/PMS and never exceeded this number (Table 2, Fig. 8A–C). In contrast, one of two examined adult males had three AC spigots on each PMS (Fig. 8D), with the distal end of the base of the late AC spigot narrower than those of the pioneer AC spigots (Fig. 8E). However, no difference was apparent in the width of the openings of late versus pioneer AC spigots (Fig. 8E) and, moreover, the other adult male only had two AC spigots on each PMS (Table 2).

Figure 8.

PMS of 4^th to 6^th instars of Australomimetus djuka. All unlabeled arrows point to non-T-AAC spigots. A, B. Juveniles. C, D. Adults. Same CY spigot (C) and closest AC spigot (unlabeled arrow) shown from two perspectives in (A, C); inset in (B) likewise from same specimen used to produce larger image, but shows lateral view of CY and AC spigots from opposite PMS. Note ontogenetic changes in CY spigot width relative to nearest AC spigot. E.AC spigots from (D) magnified, the distal ends of their shafts further magnified in insets: no obvious difference in diameters of openings, given in nm [see Materials and methods (‘SEM of spinnerets…’)], between ‘late’ AC spigot (L) and ‘pioneer’ AC spigots (P) [see Results (‘PMS-Sexual and geographic…’)]. A, B inset, C. Left PMS. B main, D, E. Right PMS (image flipped). 1°, 1° MiA spigot; C, CY spigot; L, ‘late’ AC spigot; P, ‘pioneer’ AC spigot. Scale bars: A-D 20 µm; E 10 µm. See also Abbreviations and terminology (‘Spinning apparatus abbreviations’) and Results (Convention 3 in ‘Conventions applied…’).

The PMSAC data also suggested a difference among males of A. spinosus by geography. Again, it will be necessary to examine more specimens to better evaluate this difference, but the data obtained thus far were consistent in indicating that late stadium males from eastern Australia have more AC spigots/PMS than males from Western Australia (Table 3).

Observations pertaining to PMSCY are presented below.

As with PMS-associated AC, the two species differed during the earliest stadia with respect to the number of non-T-AAC spigots/PLS: two in 1^st and 2^nd instars of A. djuka (Fig. 9A, B), increasing to three for the 3^rd stadium (Fig. 9C), as compared with three already present in 1^st instars of A. spinosus (Fig. 9D, Table 2). However, a more significant difference between the two species emerged as early as the 2^nd stadium, though more often the 3^rd, with the appearance of T-AAC spigots in A. spinosus (Table 3). In both species there were modest increases in the number of AC spigots/PLS over a spider’s ontogeny, but in A. djuka these increases consisted solely of additional non-T-AAC spigots (Table 2, Fig. 11). In A. spinosus, both non-T-A and T-AAC spigots were added to the PLS spinning field, with the non-T-AAC spigots forming a central column in the field, albeit with a slight-to-pronounced anterior displacement at the proximal end of the column (e.g., Figs 2B, 9I) and/or with the second spigot from the proximal end displaced posteriorly (e.g., Figs 9L, 10C). T-AAC spigots occurred both anterior and posterior to the non-T-AAC spigot column (Fig. 2B). Beginning with the ecdysis between the 2^nd and 3^rd stadia, the number of non-T-AAC spigots increased from one stadium to the next in both species: in A. spinosus, this increase was invariably one/PLS (Table 3), while in A. djuka the increase was variable with, in some instances, at least two such spigots apparently added to a PLS following a molt (Table 2, also cf. Fig. 11B–E, all 5^th instars). Intra-stadium variability in AC spigot number in A. spinosus was confined to T-AAC spigots (Table 3). In both species, angularities in the column formed by the non-T-AAC spigots could be present (e.g., Figs 10C, 11A) or absent (e.g., Figs 2B, 11B), but were more often present in A. djuka.

Figure 9.

PLS of 1^st to 3^rd instars of Australomimetus djuka and 1^st to 4^th instars of Australomimetus spinosus. A–C. A. djuka. D–L. A. spinosus. All spigots in (A–E, G) are non-T-AAC spigots. AC spigots between two white curves in (F, H, I, L) are non-T-A; remainder are T-A. E, F.AC spigot complements in 2^nd instars of A. spinosus: three non-T-A (E) (88%, N = 26 PLS); three non-T-A, one posterior T-A (F) (12%). G–I.AC spigot complements in 3^rd instars of A. spinosus: four non-T-A (G) (13%, N = 30 PLS); four non-T-A, one posterior T-A (H) (63%); four non-T-A, one anterior T-A (not shown) (3%); four non-T-A, one posterior T-A, one anterior T-A (I) (20%). One posterior AC tartipore (G, magnified in inset) observed in 7% of 3^rd stadium PLS; the rest had none. H.CY spigots (C) first appeared in 3^rd instars of A. spinosus. J, K.AC spigots on PLS from different A. spinosus specimens, showing larger diameter openings on T-AAC spigots (arrows) compared with neighboring non-T-AAC spigots [see Results (‘PLS’)]; measurements given in nm [see Materials and methods (‘SEM of spinnerets…’)]. L.PLS containing possible phylogenetic nubbin, adjacent to putative sensillum (S), in position consistent with araneoid AG-FL spigots. Lower inset, same nubbin (N) and sensillum magnified, anterior view; upper inset, sensillum from opposite PLS, lacking a nubbin; circles indicate (indistinct) raised-rim pores that are part of sensilla. More distinct pores can be seen in Townley and Tillinghast (2009) figs 8I, 9D, 13K, 20B, 22C and Townley et al. (2013) figs 2G, 4G. Unlabeled arrows point to AC tartipores, one anterior, one posterior, indicating an AC spigot arrangement during preceding (3^rd) stadium like that shown in (I). D, E, G, I–K, L upper inset. Left PLS. A–C, F, H, L main, L lower inset. Right PLS (image flipped). C, CY spigot; S, putative sensillum. Scale bars: A-I, L 20 µm; J, K 5 µm. See also Abbreviations and terminology (‘Spinning apparatus abbreviations’) and Results (Convention 3 in ‘Conventions applied…’).

Most 2^nd stadium PLS of A. spinosus (23 of 26 PLS from 13 2^nd instars) had the same spinning complement as 1^st stadium PLS: three non-T-AAC spigots (Fig. 9E). But the other three PLS (two from the same set of spinnerets) had, in addition, a single T-AAC spigot, consistently positioned posterior to the non-T-AAC spigot column and more distal than two of the three non-T-AAC spigots (Table 3, Fig. 9F). The identification of this AC spigot as T-A implies that we observed a similarly-positioned AC tartipore in a similarly low percentage of 3^rd stadium PLS. Indeed, 28 of 30 PLS from 15 3^rd instars contained no AC tartipores, but the other two PLS (on different sets of spinnerets) each had one AC tartipore, posterior to the non-T-AAC spigot column and more distal than two of the four non-T-AAC spigots (Table 3, Fig. 9G). This AC tartipore’s position indicated that the spigot serving the newest (fourth) non-T-AAC was the most distal one in the column. The two exceptional 3^rd stadium PLS departed further from the norm by being two of only four 3^rd stadium PLS containing no T-AAC spigots (Fig. 9G). The other 26 PLS had one (Fig. 9H) or two (Fig. 9I) T-AAC spigots (Tables 2, 3).

Consistent with the prevalence of one or two T-AAC spigots on 3^rd stadium PLS, one (Fig. 10A, yellow) and two (Fig. 10B, yellow) AC tartipores were found on the right and left PLS, respectively, of the old exoskeleton from the proecdysial 4^th stadium male A. spinosus [see Results (Convention 5 in ‘Conventions applied…’)]. It appeared that the AC accommodated by these tartipores, functional during the 3^rd stadium, would have been functional again during the 5^th stadium, as evidenced by AC spigots in essentially the same locations on the new exoskeleton of this individual (Fig. 10C, D, yellow). Spigots of T-AAC functional during the 4^th stadium are shown in blue in Fig. 10A, B and the tartipores accommodating them during this proecdysis are likewise shown in blue in Fig. 10C, D, respectively. Spigots of T-AAC that would have functioned for the first time during the 5^th stadium are shown in green in Fig. 10C, D. Though we observed intra-stadium variation in the number of T-AAC associated with a PLS (Tables 2, 3), including within the same individual (Fig. 10A, B), consistencies in the positions taken by their spigots were also often observed. For example, the posterior T-AAC spigot common to Fig. 9F, H, I occurred in essentially the same location relative to non-T-AAC spigots; likewise for single anterior and posterior T-AAC spigots common to Figs 9I, 10A, B. The distribution of T-AAC spigots, anterior and posterior to the non-T-AAC spigot column, also tended to be consistent. For example, there were 12 4^th stadium PLS that had three T-AAC spigots (as in Fig. 10B): these were invariably distributed one anterior and two posterior. And all eight 5^th stadium PLS with four T-AAC spigots (as in Fig. 10C, D) had them distributed two anterior and two posterior, while all six 5^th stadium PLS with five T-AAC spigots had them distributed two anterior and three posterior.

Figure 10.

PLS of 4^th to 6^th instars of Australomimetus spinosus. A–D.PLS from one juvenile specimen, well into proecdysis. Color coding explained in Results (Convention 5 in ‘Conventions applied…’). A, B. Old exoskeleton. C, D. New exoskeleton directly beneath (A, B), respectively. One of two AC tartipores (yellow) in (B) largely obscured (unlabeled arrow). E, F. Adult, AC spigots between two white curves are non-T-A; the rest, anterior and posterior, are T-A. AC spigots in boxed region of (E) magnified in (F), showing larger diameter openings in T-AAC spigots compared with neighboring non-T-AAC spigots [see Results (‘PLS’)]; measurements given in nm [see Materials and methods (‘SEM of spinnerets…’)]. G-I. Ontogenetic changes in CY spigot (C) morphology; medial, high tilt view. Note changes in width relative to AC spigot on right. (E, F, I) all from same PLS. B, D–F, I. Left PLS. A, C, G, H. Right PLS (image flipped). C, CY spigot; S, putative sensillum. Scale bars: A-E 20 µm; F-I 10 µm. See also Abbreviations and terminology (‘Spinning apparatus abbreviations’) and Results (Convention 3 in ‘Conventions applied…’).

Positions of T-AAC spigots and tartipores, relative to non-T-AAC spigots, in Fig. 10A–D again indicated that the addition of one non-T-AAC spigot to each PLS with each molt (starting with the 3^rd molt) in A. spinosus occurs at the distal end of the non-T-AAC spigot column. Thus, in Fig. 10C, D, the spigot of the newest non-T-AAC, which would have functioned for the first time during the 5^th stadium, is shown in purple (fuchsia) at the distal end of this column, while spigots of the five older non-T-AAC, functional at least by the 4^th stadium and again during the 5^th stadium, are shown in red in Fig. 10A–D.

Morphological differences between T-A and non-T-AAC spigots in A. spinosus were modest: the former sometimes had narrower bases than the latter, especially during earlier stadia (e.g., Fig. 9F, J). More consistent was a difference in the width of their openings, with those of T-AAC spigots wider. Absolute values of aperture diameters varied enough among different specimens, even within the same stadium, that diameters of non-T-AAC in one specimen sometimes overlapped with diameters of T-AAC in another specimen (cf. Fig. 9J, K), but within a specimen the T-AAC spigot openings were usually wider. Three examples are presented in Figs 9J, K, 10E, F.

Geographical differences between eastern and western populations of A. spinosus with respect to PLS were also modest and, as with geographical differences noted above for PMSAC, will need to be confirmed or negated by examining more material. The data currently available indicate that, from as early as the 3^rd stadium (based on AC tartipore counts in 4^th instars), eastern A. spinosus have higher mean counts of T-AAC than their western counterparts (Table 3).

No spigots of aggregate or flagelliform silk glands (AG, FL) were observed on any of the 22 A. djukaPLS and 126 A. spinosusPLS that were examined. However, a possible phylogenetic nubbin was present on the right PLS of one 4^th stadium female A. spinosus, in a position consistent with the AG-FL triad of typical araneoids (Fig. 9L). It was adjacent to a putative sensillum. Such sensilla, roughly round to oblong and containing one (not always well-defined) raised-rim pore, could be found in essentially the same location on most of the examined PLS, but with no adjacent nubbin. These are indicated in several of the micrographs shown in Figs 9–11 (labeled ‘S’).

Observations pertaining to PLSCY are presented below.

Figure 11.

PLS of 4^th to 6^th instars of Australomimetus djuka. All spigots other than CY spigots (C) are non-T-AAC spigots. A–C. Juveniles. D–F. Adults. B–E. Unlike A. spinosus, variation observed in non-T-AAC spigot number within same stadium and state of maturity. CY spigot and nearby AC spigots shown from two perspectives in (A, C, F). B, C, E, F. Left PLS. A, D. Right PLS (image flipped). C, CY spigot; S, putative sensillum. All scale bars 20 µm. See also Abbreviations and terminology (‘Spinning apparatus abbreviations’) and Results (Convention 3 in ‘Conventions applied…’).

In both species, a full complement of CY spigots, restricted to females, consisted of four CY spigots; one on each PMS and PLS.

In A. spinosus, CY spigots were first seen in 3^rd instars (Figs 6C, 9H). However, only one of five 3^rd stadium females had the full complement of CY spigots; the other four females had one to three CY spigots, with those on PMS especially prone to being absent (Table 2). Given the occurrence of one 3^rd stadium female with only one CY spigot (on the right PLS), there may exist 3^rd stadium females that take a paucity of CY spigots one spigot further and lack CY spigots entirely. If so, then one or more of the 10 3^rd stadium males included in Tables 2 and 3 may actually have been female since we relied primarily on the presence or absence of CY spigots to distinguish the sexes. Six of these 10 were intact specimens with subtly swollen palps also indicating they were males; the other four, however, were exuvia on which palps were not examined and were lost during spinneret preparation. All 4^th stadium and later females had the full complement of CY spigots (Table 2).

In A. djuka, because of the small number of early juveniles examined, we can only say that CY spigots are present at least by the 4^th stadium: the only 4^th instar examined had the full complement of CY spigots (Figs 8A, 11A), as did later females (Table 2). The only 3^rd instar examined lacked CY spigots, and its palps, not even subtly swollen, indicated it was a female. However, since we are not certain of the sex, no conclusions can be drawn regarding the occurrence of CY spigots in 3^rd stadium females of A. djuka.

In both species, increases in widths of CY spigots (bases and shafts) over successive stadia were disproportionately large compared with increases in nearby AC spigots [cf. Figs 6C (inset), 7H–J; Fig. 8A–C (insets); Fig. 10G–I; Fig. 11A, C, F (insets)]. The single largest increase occurred in shafts of CY spigots with passage from the penultimate to the adult stadium. The wide opening characteristic of mimetine CY spigots was restricted to adult females and the deep incisions present in CY spigot shafts of some mimetines were absent in both species (Figs 7C, J, 8C, 10E, I, 11F). However, shafts differed between the two species in two respects: 1) those of A. djuka were smoother, marked only by fine longitudinal striations (Figs 8C, 11F), while those of A. spinosus occupied a middle ground in terms of incising, having fine striations but also what might be described as modest incisions (Figs 7C, J, 10E, I), and 2) viewed from the side, A. spinosus shafts exhibited the domed, convex shape typical of mimetine CY spigots (Figs 7J, 10I) while A. djuka shafts were a little less rotund [Figs 8C, 11F (insets)].

The presence of PLSAC tartipores in A. spinosus (Fig. 10), which signify the presence of T-AAC and an ability to draw AC silk during proecdysis, extends the range of PLS spinning field character states known for the genus Australomimetus. AC tartipores on the PLS have been apparent in five species of Mimetus (Platnick and Shadab 1993, Griswold et al. 2005, Townley and Tillinghast 2009), two species of Phobetinus (Townley et al. 2013), and two species of the newly described mimetid genus Anansi (Benavides et al. 2016), but in none of seven species of Ero (Platnick and Shadab 1993, Schütt 2000, Townley and Tillinghast 2009) or five earlier-examined species of Australomimetus (Harms and Harvey 2009b, Townley and Tillinghast 2009). The latter included A. aurioculatus, A. diabolicus, A. maculosus, A. pseudomaculosus, and A. tasmaniensis, to which we can now add A. djuka (Fig. 11). PLSAC tartipores can also be seen in published PLS scans from at least six species of Gelanor which is the sister-lineage to all other mimetids, including Australomimetus (Platnick and Shadab 1993, Benavides and Hormiga 2016, Benavides et al. 2016). They are also present in many members of related spider families such as Tetragnathidae (Hormiga et al. 1995, Yoshida 1999, Álvarez-Padilla 2007, Álvarez-Padilla and Hormiga 2011) and Arkyidae (Platnick and Shadab 1993). Indeed, on the phylogenies presented in Wheeler et al. (2016) and Griswold and Ramírez (2017), PLST-AAC are distributed, with losses, over a clade (CY spigot clade) that includes all araneomorphs except the Synspermiata, Filistatidae, and Hypochilidae (Platnick et al. 1991, Griswold et al. 2005, Ramírez 2014). Hence this character is symplesiomorphic for the family Mimetidae, with frequent losses in some genera and species being a secondary trait. Why the character has been lost independently many times within the family is unclear and further study is needed to clarify what adaptive reasons might account for the loss. The presence of PLST-AAC in Mimetus and Phobetinus, but not Ero or (up to now) Australomimetus, was one of the characters indicating especially close affinity between Phobetinus and Mimetus (Townley et al. 2013). Our observations on A. spinosus now make this distinction less sharp. In all previously studied mimetids, if PLSAC tartipores were present, the basic arrangement on the PLS was for the non-T-AAC spigots to form a column roughly parallel to the long axis of the PLS, with all T-AAC spigots posterior to this column. A. spinosus is thus far unique, not only in being the only Australomimetus species found with T-AAC, but in being the only mimetid with T-AAC spigots forming two groups on each PLS, posterior and anterior to the non-T-AAC spigot column (Fig. 2B). Again, considering that PLST-AAC occur in several mimetid genera and are widespread among araneomorphs, this character may have relatively little phylogenetic signal and ecology/biology of the species may account for the differences observed, with follow-up studies needed to clarify why some species have this trait and others do not.

Another variable character among different species of Australomimetus is the presence or absence of 2° MiA, as revealed by the presence/absence of a 2° MiA tartipore and either a 2° MiA spigot (juveniles) or an ontogenetic 2° MiA nubbin (adults) on each PMS. Previously, only one of the five aforementioned Australomimetus species was found to lack 2° MiA: A. tasmaniensis. This study clearly places A. spinosus in the larger contingent possessing 2° MiA (Figs 6A–D, 7), while A. djuka is now the second species known without 2° MiA (Figs 6E–H, 8) and thus unable to draw MiA silk during proecdysis. It is noteworthy that the two species, A. tasmaniensis and A. djuka, constitute the tasmaniensis-group in the cladistic analysis of Harms and Harvey (2009b), sister to all other Australomimetus species included in that analysis. Possession of 2° MiA has been considered the plesiomorphic condition for araneoids on phylogenetic grounds (Forster et al. 1990, Scharff and Coddington 1997, Griswold et al. 1998), which remains valid despite substantial changes proposed in higher-level spider phylogeny in recent years (Bond et al. 2014, Fernández et al. 2014, Dimitrov et al. 2016, Garrison et al. 2016, Wheeler et al. 2016). The consistent occurrence of an apparent phylogenetic 2° MiA nubbin on the PMS of 1^st instars of A. djuka corroborates this view, clearly indicating the absence of 2° MiA in this species represents a secondary loss. This nubbin was located precisely where a 2° MiA spigot occurs in species that have 2° MiA (cf. Fig. 6A, E). Presumably, secondary loss also explains the absence of 2° MiA in five of seven examined Ero species (Platnick and Shadab 1993, Townley and Tillinghast 2009), whereas all examined species of Mimetus, Phobetinus, Anansi, and, most notably, the basal mimetid genus Gelanor have retained 2° MiA (Platnick and Shadab 1993, Griswold et al. 2005, Townley and Tillinghast 2009, Townley et al. 2013, Benavides and Hormiga 2016, Benavides et al. 2016).

Differences among species of Australomimetus with respect to these two characters, PLST-AAC and 2° MiA, suggest that some species have greater silk needs during proecdysis, and ecdysis (Dolejš et al. 2014), than others, but it remains to be determined why this might be so.

Figures 3–5 document increases in numbers of PI spigots during the ontogenies of A. spinosus and A. djuka. As observed in Mimetus and the araneid Araneus (Townley and Tillinghast 2009), these increases likely consisted solely of additional T-API spigots. The number of non-T-API spigots appeared to remain constant at two per ALS throughout development, starting from the 1^st stadium. In part, the evidence for constancy in non-T-API number came from the pooled PI data in Table 2, though admittedly, because of intra-stadium variation in numbers of T-API spigots, comparing mean PI spigot counts at one stadium with PI tartipore counts at the next stadium did not provide unequivocal evidence of such constancy. However, for those comparisons where there was no variation (1^st to 2^nd stadia in both species) or where we had larger numbers of observations (2^nd to 3^rd stadia and 3^rd to 4^th stadia in A. spinosus), a constant two non-T-API spigots/ALS was indicated.

Further evidence came from a 4^th instar male A. spinosus that was well into proecdysis when it was preserved [see Results (Convention 5 in ‘Conventions applied…’)], allowing us to examine spinnerets on both the old (4^th instar) and newly developed (pharate 5^th instar) exoskeletons. This specimen was the source of all colorized images in Figs 4, 7, and 10. As demonstrated in earlier studies (Townley and Tillinghast 2009, Dolejš et al. 2014), comparing the same spinnerets in the same individual over two or more consecutive stadia is the most reliable way to distinguish T-A and non-T-A silk glands. Matching PI spigots on the old exoskeleton (Fig. 4A) with PI tartipores on the new exoskeleton (Fig. 4B) confirmed that all but two PI emptying on the right ALS were T-A during proecdysis in the 4^th stadium. Blue T-API spigots in Fig. 4A correspond to blue PI tartipores in Fig. 4B; red PI spigots in Fig. 4A, B serve the two non-T-API associated with this ALS. The consistent positions occupied by the two non-T-API spigots were essentially the same as previously observed in Mimetus and Araneus (Townley and Tillinghast 2009: figs 5, 7).

To date, a pair of non-T-API associated with each ALS have only been indicated in three araneoid genera: Araneus, Mimetus, and now Australomimetus. In contrast, in the lycosids that have so far been examined, all PI are T-A (Dolejš et al. 2014, data for two consecutive stadia in one female Pardosa also noted in Townley and Tillinghast 2009: 376).

Dissections of ecribellate orb-web builders, especially Araneus, at various points in the molt-intermolt cycle, have demonstrated that 1° ampullate silk glands (1° MaA and 1° MiA), being non-T-A, are re-used during each successive stadium (Townley et al. 1991, 1993, Townley 1993). 2° Ampullate silk glands (2° MaA and 2° MiA), being T-A, are also re-used, but only during alternating stadia; i.e., either during even-numbered stadia for one pair each of 2° MaA and 2° MiA, or during odd-numbered stadia for the other single pairs of 2° MaA and 2° MiA. Both 1° and 2° ampullate silk glands undergo dramatic re-modeling synchronized to the molt-intermolt cycle.

Because ampullate silk glands are relatively large, and few and constant in number across juvenile stadia, gross morphological changes can be monitored by dissection over the molt-intermolt cycle. PI and AC, on the other hand, generally occur as morphological multiples (Coddington 1989) and tend to be smaller, more numerous silk glands (e.g., Kokociński 1968, Mikulska 1969, Wąsowska 1973, Mikulska and Wiśniewski 1979), less amenable to such individual monitoring by dissection. This, and the increases in AC and T-API numbers that occur during development, made it unclear whether individual AC and PI are also used during multiple stadia or only one (Townley and Tillinghast 2009). After all, given that some new PI and AC are added with most molts, might it be that the entire complement of PI and/or AC used during a given stadium (at least the T-A ones) is new and each of these silk glands functions only during one stadium? External observations made on spinnerets to date suggest the answer is no. Instead, like ampullate silk glands, once formed and functional, non-T-AAC/PI appear to function during every subsequent stadium while T-AAC/PI function during every other subsequent stadium. Dolejš et al. (2014), studying spiders of the family Lycosidae, provided persuasive evidence that relative arrangements of spigots/tartipores serving older PI/AC are to a large extent maintained during ontogeny and may, in many instances, be discerned at later stadia, this despite spigots serving newer PI and AC being added with each molt. The same conservation in positions of older PI and AC spigots/tartipores over multiple stadia was also indicated in the current study. In the proecdysial 4^th instar A. spinosus, PI tartipores on the old exoskeleton in Fig. 4A (yellow) indicated the approximate arrangement of T-API spigots that must have existed on the right ALS during the 3^rd stadium. The T-API spigots on the new exoskeleton of the pharate 5^th instar in Fig. 4B included a group (yellow) in essentially the same arrangement. This apparently conserved PI spigot arrangement during the 3^rd and (upcoming) 5^th stadia suggests that the same individual T-API that functioned during the 3^rd stadium were about to function again during the 5^th stadium. T-API spigot complements at the 3^rd and 5^th stadia thereby differed only in that the latter also included a group of T-API that would have functioned for the first, and last, time during the 5^th stadium (Fig. 4B, green): last only because this male would not have had a 7^th stadium; adulthood would have been attained in the 6^th. The consistent positions (relative to T-API spigots/tartipores) of the pair of non-T-API spigots on the old and new exoskeletons (Fig. 4A, B, red) indicated that the same individual non-T-API that functioned during the 4^th stadium would have functioned again during the 5^th stadium. Note that this pair of non-T-API spigots on each ALS are examples of PI spigots that are morphological singulars since each spigot of the pair can be individually specified: e.g., the anterior non-T-API spigot (Coddington 1989).

In the lycosids they studied, Dolejš et al. (2014) noted that new PI/AC spigots tended to be added centrifugally, at the periphery of spinning fields. This is likewise evident in Fig. 4B, where the new PI spigots (green) encircle spigots of the older PI (yellow) that had functioned during the 3^rd stadium. We therefore anticipate that, internally, when new PI corpora form, it is at the periphery of the cluster of older PI corpora.

In A. spinosus, conserved arrangements of AC spigots on the PLS during development largely mirrored those of PI spigots on the ALS. Thus, AC tartipores on the old exoskeleton in Fig. 10A, B (yellow) indicated the positions taken by T-AAC spigots during the preceding 3^rd stadium. In the pharate 5^th instar, T-AAC spigots (yellow) again occupied these positions (Fig. 10C, D, respectively), indicating that the same T-AAC that functioned during the 3^rd stadium would have returned to use during the 5^th stadium. Along with these returning T-AAC, PLS during the 5^th stadium would have had three (Fig. 10C, green) or two (Fig. 10D, green) new T-AAC available for use as well. One departure from the ALSPI condition concerned the non-T-AAC associated with the PLS (Fig. 10A–D, red + fuchsia) since their number was not constant from stadium to stadium, as it was for the two non-T-API (Fig. 4A, B, red). Instead, from the 2^nd stadium forward, the number of non-T-AAC increased by one on each PLS from one stadium to the next (Table 3), the spigot serving the new non-T-AAC added to the distal end of the non-T-AAC spigot column (Fig. 10C, D, fuchsia).

Though a given T-AAC or PI functions only during alternate stadia, there are indications, particularly from the PLST-AAC in A. spinosus, that these silk glands have some latitude with respect to when they “enter the work force”. That is, an individually specifiable T-AAC may function during odd-numbered stadia in some specimens, but function during even-numbered stadia in other specimens. Consider the following: most (88%, N = 26 PLS) 2^nd stadium PLS had no T-AAC spigots (Fig. 9E), but a minority (12%) had one posterior T-AAC spigot (Fig. 9F). This one T-AAC was responsible for an AC tartipore in the expected location (Fig. 9G), also in a minority (7%, N = 30 PLS) of 3^rd stadium PLS. These PLS with one AC tartipore had no T-AAC spigots even though most (87%) 3^rd stadium PLS had one (Fig. 9H) or two (Fig. 9I) T-AAC spigots, but no AC tartipores. This suggests that the single posterior T-AAC spigot seen in Figs 9F and 9H, consistently located, represents the same individually specifiable AC, sometimes functional from the 2^nd stadium and then again during later even-numbered stadia, but more often functional from the 3^rd stadium and then again during later odd-numbered stadia. If it functions, one might say prematurely, during the 2^nd stadium, it is not available for use during the 3^rd stadium, with the result that 3^rd stadium PLS with an AC tartipore tend to not have T-AAC spigots, while those without an AC tartipore tend to have at least one T-AAC spigot.

Adult males of A. spinosus and A. djuka did not exhibit a pair of spigots on each ALS serving modified PI (MoPI). Such spigots, located near the 2° MaA tartipore, have been found exclusively in adult males of some mimetid species. They are conspicuously different from other PI spigots, with wider shafts and openings, and segregated from the other PI spigots. It is unknown what purpose MoPI serve, but their restriction to adult males advocates for investigation of silken structures specific to courtship and mating (Bristowe 1958). The most parsimonious hypothesis regarding MoPI origin is that they are simply the pair of non-T-API associated with each ALS [see Discussion (‘Constancy in the number…’)], in use since the first stadium, but more dramatically transformed during their final proecdysial re-modeling (Townley and Tillinghast 2009, Townley et al. 2013). We cannot say with certainty that MoPI are transformed non-T-API because, in those mimetids in which MoPI form, we have not been certain that more typical non-T-API are not present concurrently. In part, this is because definitive proof of T-A or non-T-A status during one stadium resides in the exoskeleton from the next stadium, and for adults there is no next stadium. Also, because of ontogenetic increases in numbers of T-API, attempts to use conserved PI spigot/tartipore arrangements from earlier stadia [noted in Discussion (‘Recurrent use…’)] to make this determination have so far produced ambiguous results.

MoPI spigots were first described in two North American and one South American species of Mimetus (Townley and Tillinghast 2009, see also Townley et al. 2013: fig. 1B–D), and apparent in a third North American species (Griswold et al. 2005: fig. 26B). Townley and Tillinghast (2009) suggested they may be a synapomorphy for Mimetus. This was initially refuted when they were subsequently observed in adult males of the type species of Phobetinus (Townley et al. 2013). However, one consequence of the phylogenetic analyses of Benavides and Hormiga will be that Phobetinus becomes a junior synonym of Mimetus (Benavides et al. 2016). The possibility that MoPI spigots may be a synapomorphy for a relimited Mimetus that, among other changes (Benavides et al. 2016), incorporates Phobetinus (Townley et al. 2013) was also refuted when they were later observed in the most basal mimetid genus Gelanor (adult males of three species, Benavides and Hormiga 2016). Thus, MoPI spigots may be a synapomorphy for the entire family Mimetidae, again with secondary losses having occurred in individual species or even genera.

MoPI spigots like those in Mimetus, Phobetinus, and Gelanor have not been evident in those Ero or Australomimetus species examined to date (Harms and Harvey 2009b, Townley and Tillinghast 2009). However, adult males of two of five examined Ero species have been found with what have been referred to as ‘subtle MoPI spigots’, with modestly wider openings than other PI spigots, but not substantially wider shafts and not set apart from other PI spigots (Townley and Tillinghast 2009). Like MoPI and non-T-API spigots, they number two per ALS. Subtle MoPI spigots were not found in A. aurioculatus, A. diabolicus, or A. tasmaniensis (Townley and Tillinghast 2009), but Townley et al. (2013) noted their presence in A. sydneyensis. The morphology and positions of subtle MoPI spigots are so similar to those of non-T-API spigots that a non-T-API origin for subtle MoPI seems inescapable. This, in turn, supports a non-T-API origin for MoPI as well, with subtle MoPI seemingly exhibiting a degree of transformation during the last proecdysis intermediate between the high extreme of MoPI and the low extreme of typical non-T-API. Presumably, internal investigations of the PI will make MoPI and subtle MoPI origins clear. In any case, A. sydneyensis is currently unique among Australomimetus, since A. spinosus and A. djuka did not exhibit even subtle MoPI spigots in adult males (Figs 4D, E, 5B, C).

CY, also known as tubuliform silk glands, are used by adult females in the construction of egg sacs. The two Australomimetus species examined in this study exhibited the unusually large, wide-aperture CY spigots typical of adult female mimetines (Platnick and Shadab 1993), one per PMS and PLS, unlike the two per PLS that are present in gelanorines (Platnick and Shadab 1993, Benavides and Hormiga 2016) and considered plesiomorphic for araneoids (Griswold et al. 1998). CY spigots were also present in juvenile females, as early as the 3^rd stadium in A. spinosus, which agrees with findings in two Mimetus species (Townley and Tillinghast 2009). We are not, however, aware of any evidence that they are used in juveniles. To the contrary, histological observations indicate that, at least in general, CY become functional only in adults concurrent with vitellogenesis in the ovaries (Richter 1970, Kovoor 1977, 1987). If juvenile mimetine CY were to produce silk fibers, they would presumably be much thinner than those drawn from adult CY spigots since juvenile CY spigots lack the wide shafts and openings of adults. It was observed in A. spinosus that the full complement of four CY spigots was not present in a large majority of 3^rd stadium females. Similarly, in Araneus cavaticus, CY spigots first appear in 4^th stadium females, but often not with the full complement of six spigots present during later stadia (Townley and Tillinghast 2009). Müller and Westheide (1993) hypothesized that the presence of CY spigots in juveniles indicates that their female ancestors matured at earlier stadia. If true, this would suggest that female ancestors of Australomimetus matured as early as the 3^rd stadium, which is at least a possibility among small spiders (Gertsch 1979).

The deep incising typical of CY spigot shafts in Ero and Mimetus (Platnick and Shadab 1993, Schütt 2000, Griswold et al. 2005, Townley and Tillinghast 2009, Townley et al. 2013), but not Phobetinus (Townley et al. 2013) or previously examined Australomimetus (Harms and Harvey 2009a, b, Townley and Tillinghast 2009), was also not evident in A. spinosus or A. djuka. There were, however, more minor differences between CY spigot shafts in the two species, including shallow incising in A. spinosus that was lacking in A. djuka [see Results (‘Cylindrical silk gland…’)].

Remarkably, though the newly described mimetid genus Anansi nests within the subfamily Mimetinae, scans of their spinnerets (Benavides et al. 2016) demonstrate that adult females have more conventional CY spigots and not the enlarged, rotund, wide-aperture CY spigots of most mimetines. Moreover, while at least one species has a single CY spigot on each PLS, typical of mimetines, at least two other species have no CY spigot on the PLS (Benavides et al. 2016). The latter retain only the single CY spigot on each PMS [fig. 17G in Benavides et al. (2016) has two spigots on a PMS labeled CY, but the lower one, given its position near the 2° MiA nubbin and tartipore, is presumably the 1° MiA spigot despite its morphology being similar to the authentic CY spigot]. These departures from the typical mimetine condition may be attributable to maternal care behavior observed in the genus (Benavides et al. 2016). Females carry eggs and postembryos with their chelicerae, and photos of this behavior indicate minimal wrapping with silk. This contrasts with the substantial egg sacs of most other mimetines (Guarisco and Mott 1990, Guarisco 2001, Thaler et al. 2004, Harms and Harvey 2009b, Kůrka et al. 2015, Killick 2016) and lack of maternal care (Bristowe 1958). Benavides et al. (2016) also observed maternal care and apparently insubstantial egg sacs in two neotropical mimetines currently placed in Mimetus. It will be of interest to determine the status of their CY spigots.

With the exception of a single PMS, there were two AC spigots on each PMS in four adult female A. spinosus. In contrast, five adult males had 3-5 AC spigots on each PMS and one juvenile male from eastern Australia had three per PMS (Table 3). Published data for A. aurioculatus indicate that essentially the same sexual dimorphism also applies to that species (Townley et al. 2013, see fig. 18G, H in Townley and Tillinghast 2009). Two adult female A. aurioculatus had two AC spigots per PMS while two adult males had 4-6 AC spigots per PMS. The additional (‘late’) AC spigots in males, with wider apertures than the two ‘pioneer’ AC spigots (Fig. 7D–G), extended toward the area occupied by CY spigots in females, raising the possibility that this dimorphism may be attributable to space constraints. It is true, however, that large CY spigots have not prevented some adult female mimetines (North American Mimetus species) from fitting four AC spigots per PMS (Griswold et al. 2005: fig. 25C, Townley and Tillinghast 2009: fig. 17B). Our very limited data from A. djuka indicate a less distinct difference in PMSAC spigot number between the sexes.

The specimens of A. spinosus examined in this study included examples from Western Australia and from the eastern states of Queensland and New South Wales. Only slight differences in spinneret features were indicated between these populations, and because of the limited data, additional observations will be required to verify or refute them. These differences concerned the (never high) numbers of AC spigots, with eastern specimens tending to have more T-AAC spigots on PLS and, among older males, more non-T-AAC spigots on PMS (Table 3). There is recent genetic evidence that Western Australian A. spinosus populations are sufficiently distinct from eastern populations to define them as a separate species (Harms in prep.), but the morphology of the genitalia in both sexes is cryptic and additional studies of spigot structure may offer a possibility to delineate both candidate species on morphological grounds. Ecology certainly suggests a split into distinct species because the WA populations of A. spinosus occur on rocky outcrops whereas the eastern populations are from subtropical rainforest habitats and the observed AC differences, if consistent, may reflect ecologically-driven genetic divergence.

In most, if not all, araneomorphs, functional T-A silk glands are first available for use in 1^st instars (Townley and Tillinghast 2003, Dolejš et al. 2014). Consequently, tartipores first form and function in the exoskeleton of the pharate 2^nd instar, during proecdysis of the 1^st stadium [see Abbreviations and terminology (‘Terminology’), Fig. 1], and are therefore not observable externally until the 2^nd stadium. In 1^st instars of A. spinosus and A. djuka, a round/oval structure was seen lateral to the 2° MaA spigot on each ALS (Fig. 3A, D), near the location taken by a 2° MaA tartipore beginning in pharate 2^nd instars. A comparable structure was observed adjacent to the 2° MiA spigot on each PMS, but only in A. spinosus (Fig. 6A). Such structures have been noted before, on 1^st instars of lycosids, gnaphosids, uloborids, araneids, and mimetids (Mimetus) (Townley and Tillinghast 2003, 2009), and presumed to be ampullate silk gland tartipore primordia (called pre-tartipores in the 2003 paper) given their locations and restriction to 1^st instars. It is consistent with this identification that A. djuka, a species that lacked 2° MiA tartipores at all stadia (Figs 6G, H, 8), showed no sign of these putative primordia on PMS of 1^st instars (Fig. 6E, F).

The position taken by the structure in question on PMS of 1^st instars of A. spinosus was also consistent with a tartipore primordium identity. Relative to the 2° MiA spigot, its location varied over about a 90° arc on different 1^st stadium PMS, from directly anterior to the spigot (as in Fig. 6A) to directly lateral to the spigot. In 2^nd and later instars, the 2° MiA tartipore and spigot (or nubbin, in adults) switched positions from one stadium to the next (Figs 6B–D, 7A–D, F), such that the tartipore was posteromedial to the spigot in even-numbered instars and anterolateral to the spigot in odd-numbered instars (Table 2). Thus, a 2° MiA tartipore primordium would be expected in the anterolateral position in 1^st instars. The anterior-to-lateral position that was observed fits this prediction well.

Like other Mimetidae (Platnick and Shadab 1993, Schütt 2000, Griswold et al. 2005, Townley et al. 2013, Murphy and Roberts 2015, Benavides and Hormiga 2016), including other species of Australomimetus (Harms and Harvey 2009a, b, Townley and Tillinghast 2009), the two species studied here lacked spigots of AG and FL on the PLS. As these two silk gland types are synapomorphic for Araneoidea (Coddington 1986, Griswold et al. 1998, Schütt 2000, 2003), these absences presumably represent secondary losses associated with araneophagy and abandonment of capture-web construction (Platnick and Shadab 1993, Schütt 2000). However, it may be that vestiges of these silk glands remain in the form of occasional phylogenetic nubbins and associated sensilla. In a SEM study of 156 PLS from 1^st to 6^th instars of Mimetus, nine (6%) had a cuticular protuberance in a location consistent with an araneoid AG-FL triad (three spigots: two AG, one FL) (Townley and Tillinghast 2009). These protuberances were tentatively identified as phylogenetic nubbins, possibly of AG or FL spigots, a suggestion consistent with a growing consensus that mimetids are closely related to the araneoid family Tetragnathidae (Blackledge et al. 2009, Dimitrov and Hormiga 2011, Dimitrov et al. 2012, Bond et al. 2014, Hormiga and Griswold 2014, Benavides et al. 2016, Dimitrov et al. 2016, Garrison et al. 2016, Wheeler et al. 2016). Putative sensilla associated with these nubbins and with AG/FL spigots in typical araneoids (several indicated in Figs 9–11), and similarities between these nubbins and those occasionally seen in the same location in the araneid Cyrtophora, have been discussed previously (Townley and Tillinghast 2009, Townley et al. 2013). In this study, a total of 148 PLS from 1^st to 6^th instars of Australomimetus were examined, and only one nubbin of this type was observed (0.7%) (Fig. 9L).