Natural History of Glenognatha Emertoni (Araneae, Tetragnathidae): Mating Behavior and Sperm Release in a Haplogyne

2006. The Journal of Arachnology 34:387–398

NATURAL HISTORY OF GLENOGNATHA EMERTONI (ARANEAE, TETRAGNATHIDAE): MATING BEHAVIOR AND SPERM RELEASE IN A HAPLOGYNE

A. M. Danielson-Franc¸ois1: Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, Arizona 85721 USA.

ABSTRACT. Glenognatha emertoni (Simon 1887) is a small haplogyne orb-weaver collected near streams and dry streambeds in southern Arizona whose habits are unknown. Field observations revealed that G. emertoni are commonly found in vegetation overhanging streams and, more rarely, under streamside rocks. Mating pairs were observed on or near adult female webs. Males lack mate-guarding behavior and leave the female immediately after copulation. To examine mating behavior in a controlled setting, juveniles and adults were collected from the ﬁeld and maintained in the lab. Matings were arranged between wild-caught adults and also between laboratory-reared virgins in order to describe mating behavior and sperm release during copulation. Unlike most other orb-weaving spiders studied, the number of sperm released and overall duration of copulation are not inﬂuenced by female mating history in G. emertoni. Male G. emertoni release equivalent numbers of sperm to virgin and non-virgin females. Given this pattern of sperm release and the lack of mate-guarding behavior by males, sperm competition should be intense in this species. Based only on the numbers of sperm released by each male in the study, doubly-mated females would be expected to produce egg sacs of mixed paternity, if all else were equal. Keywords: Araneae, sexual selection, sperm competition, spermatheca, spider

Glenognatha emertoni (Simon 1887) is a from Barrows (1919) observations of Glen- small, haplogyne tetragnathine orb-weaver ognatha foxi (McCook 1894) and a single (adult body length 4.5–5 mm) known from published note by Edwards & Senske (2001) Arizona and New Mexico (Map 8, Levi 1980). on the mating behavior of Glenognatha hel- The habits of G. emertoni are unknown. The eios (Hormiga 1990), a recently discovered genus Glenognatha includes 20 named spe- species (Hormiga & Do¨bel 1990). cies from the Americas (North, Central, and Haplogyne taxa are especially intriguing to South), Africa, and the Caribbean, Galapagos, study because they provide a counter-point to and Pacific Islands, but it is possible that studies on entelegynes, arguably the largest many tropical species remain undescribed and most well-researched group of spiders. (Hormiga & Do¨bel 1990; Platnick 2006). This Haplogyny is the basal condition in the Ara- genus is a member of the sub-family Tetrag- neomorphae and haplogynes are defined, in nathinae, a haplogyne clade within the ente- part, by their simple genitalia (Wiehle 1967). legyne family Tetragnathidae (Hormiga et al. In particular, haplogyne females have cul-de- 1995). Within the tetragnathines, the mating sac spermathecae with one duct for insemi- behavior of species in the haplogyne genera nation and fertilization (Simon 1895). Ente- Tetragnatha and Pachygnatha have been legyne females have more complex conduit well-described (Gerhardt 1921, 1927; LeSar spermathecae, each with a specialized duct for & Unzicker 1978; Huber 1998; West & Toft insemination whose opening is located exter- 1999; Danielson-François 2002; Danielson- nally in the epigynum, and a separate duct for François & Bukowski 2005). However, the be- fertilization (Simon 1895). Although they are havior of species within the haplogyne genus located within the entelegyne suborder (Hor- Glenognatha is essentially unstudied apart miga et al. 1995; Griswold et al. 1998, 1999), a reversal to haplogyny has occurred within 1 Current address: Department of Ecology and Evo- the tetragnathines (Wiehle 1967), which have lutionary Biology, Rice University, P.O. Box 1892, cul-de-sac spermathecae with one duct for in- Houston, Texas 77251-1892 USA. E-mail: semination and fertilization (Danielson-Fran- [email protected] çois 2002).

387 388 THE JOURNAL OF ARACHNOLOGY

Such morphological differences in female naeus 1758) (Yoward 1996, 1998; Kaster & spermathecae might underlie the diversity of Jakob 1997; West & Toft 1999; Scha¨fer & Uhl paternity patterns in spiders, which range from 2002). In contrast, Physocyclus globosus (Tac- first- to last-male advantage—unlike the last- zanowski 1873) has a paternity pattern that male advantage common to insects, whose does not significantly differ from random spermathecae often resemble a simple cul-de- sperm mixing (Eberhard et al. 1993). Haplo- sac (Walker 1980; reviewed in Simmons & gyne paternity patterns generally exhibited a Siva-Jothy 1998; Elgar 1998; Eberhard 2004). high level of variation within a species: males To explain this diversity, Austad (1984) pro- sired from 0 to 100% of the progeny—similar posed that spermathecal morphology and ejac- to other spider paternity studies (reviewed in ulate stratification generated first-male advan- Elgar 1998). Even within an individual fe- tage in entelegynes and last-male advantage in male, paternity patterns may vary: subsequent haplogynes. Austad (1984) also argued that egg sacs from mated T. extensa females gen- haplogyne males would be indifferent to- erally revealed an anti-chronological order of wards, but that entelegyne males would prefer, mating priority; however, a series of egg sacs virgin females as mates and that haplogyne from individual females defied any such gen- males would guard mates after copulation. To eral rules (West & Toft 1999). From these date, mate-guarding, mating behavior, and pa- studies, it seems clear that paternity patterns ternity have been studied in relatively few are highly variable both within and among haplogyne taxa. haplogyne species. Post-copulatory mate-guarding should be Although many different factors certainly common in haplogyne taxa if Austad’s (1984) underlie these complex behavioral and pater- predictions hold. Of the haplogyne species nity patterns (e.g., cryptic female choice; studied thus far, co-habitation of adult pairs Eberhard 1996, 2004), differences in sperm has been found in seven pholcid taxa (Eber- release among males might explain part of the hard & Bricenõ 1983; Eberhard et al. 1993; variation. The amount of sperm transferred by Blanchong et al. 1995; Kaster & Jakob 1997). males to virgin and non-virgin females may Contrary to Austad’s hypothesis, in at least vary, as has been demonstrated in several en- one pholcid species males co-habit equally telegyne taxa in which males release little or with penultimate-molt and mature females no sperm to non-virgin females (Christenson (Eberhard et al. 1993). Yet overall, a strong & Cohn 1988; Andrade 1996; Bukowski & preference for co-habitation with penultimate- Christenson 1997; Bukowski et al. 2001). In stage females, rather than mature females, ap- contrast, the only data available for a haplo- pears to be lacking in the pholcid species stud- gyne species thus far showed that males re- ied (Eberhard et al. 1993). Of the Tetragnatha lease equivalent numbers of sperm to virgin species examined, both co-habitation and and non-virgin females (Danielson-François mate-guarding appear to be absent; males & Bukowski 2005). Data from only one spe- leave the female immediately after copulation cies prevents generalizations about haplogyne (West & Toft 1999; Danielson-François 2002; sperm release behavior: more research on hap- Danielson-François & Bukowski 2005; but logyne species is needed. see LeSar & Unzicker 1978). No information This study focused on an Arizona popula- on the co-habitation or mate-guarding behav- tion of G. emertoni from the Huachuca moun- ior of Glenognatha species is available at pre- tains, an appropriate location because the type sent. specimen for the species was also collected in Austad (1984) predicted that last-male pa- southeastern Arizona (Levi 1980). Here, I de- ternity advantage would be common among scribe the natural history and mating behavior haplogyne taxa, but empirical evidence to date of G. emertoni, while offering the second offers equivocal support. Paternity patterns demonstration of sperm release behavior in a have been examined in three pholcids and one haplogyne species. Tetragnatha species (reviewed in Elgar 1998; Eberhard 2004). A last-male biased paternity METHODS pattern has been shown for Pholcus phalan- Natural History and Collection.—Previ- gioides (Fuesslin 1775), Holocnemus pluchei ous records for Glenognatha emertoni stated (Scopoli 1763), and Tetragnatha extensa (Lin- that specimens were collected under rocks DANIELSON-FRANÇOIS—BEHAVIOR& SPERMRELEASEINA HAPLOGYNE 389 near streams or in dry streambeds in the glass aquaria (30.5 ϫ 15.0 ϫ 20.5 cm) con- southern corner of Arizona and New Mexico taining about 4 cm of water and several twigs. (Levi 1980). Over the course of this study, Females were allowed five minutes to settle approximately 175 immature, penultimate- onto their twig before a randomly chosen male molt and adult G. emertoni were collected was placed nearby. All encounters were vid- from vegetation overhanging streams in mid- eotaped with a NEC color CCD camera model elevation riparian areas in the Huachuca NX18A with attached macro lens (Micro Nik- Mountains of southeastern Arizona (31Њ28ЈN, kor Nikon 105 mm) mounted on a wooden 110Њ20ЈW). Voucher specimens were depos- platform with a Bogen ball head for camera ited in the arthropod collection of the Depart- rotation. During recording, live signal was re- ment of Entomology, University of Arizona, layed to a monitor, resulting in ϳ20ϫ mag- Tucson, Arizona 85721 USA. Of the spiders nification. Similar to other tetragnathine spe- collected, 40 penultimates were brought into cies studied, G. emertoni also use chelicerae the laboratory, and reared to adult stage in in- as holdfasts to steady the pair as the male individual 20 ml polystyrene vials from March– serts his pedipalps during mating. After cop- May 2000. Spiders were fed ad lib. daily on ulation, staged encounters were considered fruit flies (Drosophila melanogaster) and wal- terminated when the pair disengaged chelic- nut flies (Rhagoletis juglandus). The animals erae and retreated from one another for at least were observed daily, and the date of molting 15 min. Pairs have never been observed to was recorded. Adult virgins were used in resume copulation more than 15 min after staged mating experiments in the laboratory. cheliceral disengagement (Danielson-François Several wild adults were collected from pers. obs.). If mating did not occur within 15 March–May 2000 and additional wild adults min, the previous male was removed and a were collected from March–April 2001. Col- new male was introduced to the mating arena lection sites were far downstream of field ob- until mating occurred. Adult females were servation areas in order to avoid influencing mated to two adult males within fifteen min- the natural behavior of individuals at the ob- utes of each other. Males were immediately servation sites. frozen and stored at Ϫ80Њ Cforspermquan- Field observations of mating behavior were tification. Procedures for mating wild-caught recorded during the late afternoon into the ear- males to wild-caught females were similar to ly evening (15:00–22:00 h) on 31 April and those described above for mating laboratory- 2, 4, 6 May 2000 under ambient light condi- reared virgins. tions whenever possible. Occasionally, when Videotapes were later transcribed and laten- natural light was insufficient for observations, cy to cheliceral engagement (in sec), latency a flashlight also was used. When used, the to courtship after cheliceral engagement (in flashlight was not focused directly on the spi- sec), which palp was inserted (left or right), ders, but directed instead at the substrate be- duration of palp insertion (in sec), number of neath the web such that the individual was insertions, and total time spent in copula (in silhouetted against a brighter background. sec) were recorded. Palp insertion was defined Two females on webs with males nearby were as the complete insertion of the tip of the em- observed each evening. Web surfaces were bolus into the female’s gonopore that was suc- horizontal (or nearly so) and built directly ceeded by at least one full hematodochal in- above the stream, usually within five cm of flation and deflation. the water surface. Field observations of mat- Procedures for Sperm Quantification.— ing behavior were comparable to the results The methods for sperm quantification are obtained in the laboratory, suggesting that the based on a modification of the protocol of Bu- staged matings were an appropriate surrogate. kowski & Christenson (1997). Male palps Procedures for Staged Matings.—Labo- were removed at the femur under a dissecting ratory-reared virgin males were at least three microscope. Right and left palps were then la- days post-molt before being used in staged beled and separated. Each palp was placed matings to ensure they had inducted sperm into a labeled 1.5-mL polypropylene centri- into their palps. Prior to mating, virgin labo- fuge tube (Brinkman) containing 300 ␮L of a ratory-reared females that were at least seven sonication solution of 1 mL of 0.9% saline days post-molt were individually placed into and 10 ␮L of 10% Triton-X100 detergent 390 THE JOURNAL OF ARACHNOLOGY drawn from a common stock. Treatment with copulation, for each individual. The average detergent was necessary to prevent sperm ag- sperm count of the unused virgin palps was gregation and facilitate a homogeneous distri- used to estimate the amount of sperm avail- bution within a sample. Each palp was able prior to copulation in the staged matings. crushed with forceps and centrifuged at 1000 The amount of sperm available prior to cop- g for 5 min. After centrifugation, each sample ulation for wild-caught adult males was un- was ultrasonicated using a Branson sonicator known, as males may be sperm-depleted or with a 3.2 mm probe at a low level (approx- have higher reserves from re-inducting sperm imately 2% of total power output) for about multiple times. To be conservative, calcula- 20 s. Exactly 300 ␮l additional sonication so- tions of sperm release from wild-caught adult lution was added to the preparations that were males were based on the average sperm count then vortexed for approximately 10 s. A sub- of the unused palps of virgin laboratory-reared set of the homogenized sample was immedi- males. For comparison, several wild-caught ately withdrawn and placed on a hemacytom- males were allowed to mate using only one eter. Sperm were counted under a light palp (in a similar fashion as above) to directly microscope at 400ϫ magnification. All values determine how much sperm was available be- reported are estimates of total numbers of palp fore copulation and released during palp in- sperm based on linear extrapolations of actual sertion for those particular males. counts to the total sample volume. Statistical Analyses.—The Mann-Whitney Estimating Sperm Release.—To estimate U test was used to assess differences in the the amount of sperm available prior to mating, amount of palp sperm among and within vir- a subset of males were sacrificed and their gin and mated males. The Kruskal-Wallis test sperm was counted (see below). The resulting was used for comparisons between males mat- average sperm count was used as an estimate ed to virgin, non-virgin, and wild-caught fe- of the amount of sperm available prior to cop- males. Non-parametric Spearman rank corre- ulation. Sperm release is calculated by sub- lation was used to test for associations tracting the amount of sperm remaining in the between copulation duration and sperm re- palps after copulation from the amount of lease. All summary statistics of continuous sperm available prior to copulation. This is a variables are reported as X¯ Ϯ SE. conservative estimate of sperm release for males used in the mating studies. A more ac- RESULTS curate estimation would be to restrict all males Field Studies.—The following section de- in the mating studies to one palp insertion, and scribes the natural history of G. emertoni, in- compare used and unused palps for each male cluding observations on mating behavior. to directly determine sperm release. However, Habits: Overall, 112 adult, 40 penultimate- limiting males to only one palp insertion molt and 23 immature individuals were ob- would create other artifacts when examining served. During daylight hours, adult and pen- the overall length of mating, so this method ultimate individuals (respectively, n ϭ 69, n was only used for this subset of males. ϭ 40) were found near the stream under To measure sperm release, a subset of vir- leaves, rocks or overhanging vegetation, gin males were allowed to copulate and trans- whereas immatures (n ϭ 23) were found in fer sperm from only one palp to a virgin fe- miniature webs suspended in between grass male before being sacrificed for sperm stems approximately three cm above the counting (i.e., half-virgin males). After the stream. When resting on vegetation (e.g., male completed his first palp insertion (i.e., grass) diurnally, G. emertoni adopt a cryptic removed the first palp and began to position posture by extending leg pairs (1, 2) parallel his other palp for insertion), the male was im- to and wrapping leg pairs (3, 4) around a stem. mediately removed and placed at Ϫ80Њ C for Occasionally, adult females were found on the sperm counting later. These males are de- remnants of a web during the late afternoon scribed as ‘‘one-insertion’’ males. This al- on overcast days (n ϭ 7), but were more often lowed a direct comparison between the in their daytime retreats (n ϭ 19). Retreats amount of sperm in the unused and used consisted of a few silk lines attached to the palps, i.e. the amount of sperm available prior vegetation. Diurnally, adult and penultimate to copulation and the amount released during individuals were found in retreats (respective- DANIELSON-FRANÇOIS—BEHAVIOR& SPERMRELEASEINA HAPLOGYNE 391 ly, n ϭ 43, n ϭ 40). After dusk, female adults fold and, therefore, are not visible during mat- were found on webs (n ϭ 27) and male adults ing. The two complete matings observed in were either found moving between aggrega- the field were shorter in duration (n ϭ 2, total tions of females (n ϭ 5) or resting beside a duration X¯ ϭ 936 Ϯ 240 sec) than those ob- female’s web (n ϭ 11). served in the laboratory. Adults were observed rebuilding their webs Males left females immediately after mat- nightly. Typically, individuals emerged after ing in the field, suggesting that there is no dusk and began to spin webs three to five cm post-copulatory mate guarding in this species. above the water surface. On multiple occa- Yet, in the late spring when most of the mat- sions, adult females built webs inside the ings were observed, females were rarely col- webs of Tetragnatha versicolor (Walckenaer lected from their daytime retreats without at 1842), but no territorial aggressions between least one male nearby. More often, a female the two species were seen either during or af- was collected from her retreat with two or ter web construction. three males nearby (n ϭ 11, n ϭ 4, respec- Avoidance Behavior: When disturbed, G. tively). Occasionally, an adult pair was col- emertoni dropped into the water beneath and lected with up to five males in the surrounding traveled downstream, floating along with the vegetation (within 10 cm of the pair, n ϭ 4). current. To emerge from the water, they ex- It is unknown if males compete for access to tended their legs, which retarded their down- females within these groups during the day stream movement, and climbed up streamside (see Discussion). vegetation. During collection, several G. Mate-Avoidance Behavior: On one occa- emertoni traveled approximately one meter sion in the field, an unusual mate-avoidance downstream on the current, but more escaped behavior was observed between a pair of until an appropriate technique was developed: adults with interlocked chelicerae on a stem placing a mesh net underneath the individual of grass overhanging the stream. The female as collection was attempted. appeared to be resisting copulation: she held Mating Behavior: Three matings were ob- her abdomen away from the male and began served in the field. Males approached females moving backwards, pulling the male down the on webs (n ϭ 2) or resting on vegetation (n stem towards the water. Once she made con- ϭ 1, this female left before sperm transfer and tact with the stream, she released her hold on is excluded from analyses below, see Mate- the grass entirely and floated on the water, still Avoidance Behavior). Upon contact, adults held by the male’s chelicerae interlocked with immediately paired in a ventral-to-ventral her own. Seconds later, after touching the fe- mating position by interlocking their chelic- male with leg pairs (1, 2), the male disen- erae. Then males began a vibratory courtship gaged his chelicerae and during this momen- phase (n ϭ 2, latency to courtship X¯ ϭ 11 Ϯ tary release the female was quickly pulled 6 sec, courtship duration X¯ ϭ 123 Ϯ 37 sec); away by the current. Approximately one meter this phase was considered courtship because downstream, the female extended her legs to it occurs before palp insertion and sperm grasp another grass stem and emerged from transfer. During courtship the male rapidly the water. The male remained on the original flicked leg pairs (1, 2) over the female, until grass stem, skimming leg pairs (1, 2) across she exhibited an acceptance posture—curling the water surface for about 30 s before return- her abdomen towards the male—a necessary ing to a resting position higher up on the grass step to bring her gonopore within reach of his stem. Another example of this mate-avoidance palps. behavior also occurred in the laboratory. In the field, a male inserted one palp, and Egg Sacs: Egg sacs were between pairs of then inserted the other palp, using each palp moist leaves stitched tightly together with silk only once (n ϭ 2, copulation duration X¯ ϭ located in streamside litter, just above the wa- 803 Ϯ 270 sec). The insertion of pedipalps ter line. The base of the egg sac was spun onto appears to be ipsilateral, but further confir- a leaf surface, eggs were deposited, and a sec- mation of this would require freezing pairs in ond tightly woven sheet was placed over the copula to determine the placement of the palp eggs before being attached loosely to the other because the openings to the female sperma- leaf. The egg sac was distinctly flattened, its thecae are located interior to the epigastric upper surface being slightly concave. Two 392 THE JOURNAL OF ARACHNOLOGY

Table 1.—Summary of mating behavior for laboratory-reared virgin males mated to virgin and non- virgin females, and wild-caught males mated to wild-caught females in the laboratory, X¯ Ϯ SE (n).

Female type Virgin Non-Virgin Wild-Caught Latency to courtship (s) 40 Ϯ 7.6 (8) 42 Ϯ 7.9 (8) Ͻ1 (3) Courtship duration (s) 126 Ϯ 34 (8) 151 Ϯ 14 (8) 110 Ϯ 5 (3) Copulation duration (s) 1,000 Ϯ 180 (8) 700 Ϯ 110 (8) 1,734 Ϯ 400 (3) Total duration (s) 1,250 Ϯ 180 (8) 1,050 Ϯ 160 (8) 2,035 Ϯ 626 (3) sacs were collected, each of which had 20–30 their chelicerae after the first palp insertion eggs inside. The egg sacs were fairly large (3– and then immediately re-engaged chelicerae 4 mm across) relative to the adult body size and re-courted the female before inserting the of G. emertoni (4.5–5 mm). In the lab, females second palp. All wild-caught males inserted laid multiple sacs (maximum ϭ 4) ranging each palp only once. Sperm extrusion was not from 16 to 26 eggs per sac. The hatchlings observed during or after any staged mating. had a well-developed stout cephalothorax and, Mating propensity, latency to courtship, upon casual visual inspection, appeared to be courtship duration, palp insertion duration, larger than hatchlings of Tetragnatha versi- and overall mating duration were compared color (Walckenaer 1842) and Leucauge ven- for all females. Laboratory-reared males were usta (Walckenaer 1842). equally as likely to mate virgins as non-vir- Laboratory Studies.—The following sec- gins (n ϭ 18, ␹2 ϭ 1.80, P ϭ 0.18); ten virgin tion describes the experimentation on mating females were mated immediately by the first behavior and sperm release. male presented to them and eight mated with Mating Behavior: In the laboratory, just as the second male presented to them. (The two in the field, males and females interlocked singly-mated females and their mates were ex- chelicerae immediately upon contact and cluded from the analyses.) Latency to court- males began courtship less than a minute later ship (after cheliceral engagement), was not (Table 1). Vibratory courtship consisted of the significantly different for males courting vir- male flicking his legs (pairs 1, 2) dorsally over gin and non-virgin females, but was lengthier the female until she adopted an acceptance than that of wild-caught males (Mann-Whit- posture. The male then used his legs to move ney, U ϭ 29, P ϭ 0.75; Kruskal-Wallis, df ϭ the female’s abdomen closer, sometimes shak- 2, H ϭ 7.3, P ϭ 0.03, respectively; Table 1). ing her until he successfully inserted his first In contrast, the courtship duration was signif- palp. icantly longer for laboratory-reared males Upon insertion, the palp hematodochae in- courting non-virgin than virgin females, but flated and deflated repeatedly. In the staged neither was significantly different from court- matings, both laboratory-reared virgin and ship of wild-caught males (Mann-Whitney, U wild-caught males alternated palps, and in ϭ 13, P ϭ 0.05; Kruskal-Wallis, df ϭ 2, H ϭ most cases, inserted each palp only once be- 5.3, P ϭ 0.07, respectively; Table 1). When fore disengaging chelicerae. There were no all palp insertions were considered jointly, in- significant differences in the number of palps sertion duration was not different between used by virgin males when mating with virgin first and second laboratory-reared males al- and non-virgin females (Fisher’s exact test, P though there was a trend for insertions with ϭ 0.23), but two males mating non-virgin fe- non-virgins to be shorter in duration (n ϭ 16, males did insert a palp twice. In each case, the virgin X¯ ϭ 503 Ϯ 71; n ϭ 16, non-virgin X¯ palp inserted a second time was the one with ϭ 348 Ϯ 65; Mann-Whitney, U ϭ 83, P ϭ the shortest initial insertion duration. These 0.14). When only first insertions were consid- matings were otherwise as described above ered, males mated to non-virgin females had with cheliceral engagement lasting the dura- shorter insertions than those mated to virgin tion of the mating. Out of 16 matings, two females (Mann-Whitney, U ϭ 11, P ϭ 0.05); laboratory-reared males (not the same males however, when second insertions were com- from the previous case) briefly disengaged pared, there was no difference (Mann-Whit- DANIELSON-FRANÇOIS—BEHAVIOR& SPERMRELEASEINA HAPLOGYNE 393

Table 2.—Total sperm released and sperm remaining for both palps of virgin males mated to virgin and non-virgin females in the laboratory, X¯ Ϯ SE (n). *To calculate release, the sperm retained in the mated palp was subtracted from the estimate of sperm available prior to copulation.

Female type Virgin Non-Virgin Sperm released* 710,000 Ϯ 87,000 (8) 740,000 Ϯ 87,000 (8) Sperm remaining 300,000 Ϯ 110,000 (8) 200,000 Ϯ 78,000 (8)

ney, U ϭ 29, P ϭ 0.79; Table 1). Wild-caught all mating behavior and sperm release analy- males had lengthier palp insertions than lab- ses). oratory-reared males (Kruskal-Wallis, df ϭ 2, Sperm Release From Laboratory-Reared H ϭ 6.1, P ϭ 0.05; Table 1). There was no Males: All males released sperm (n ϭ 16). significant difference in total copulation du- Males mated to virgins and those mated to ration between males mated to virgin and non- non-virgins (i.e., first and second males to virgin females and those mated to wild-caught mate) did not have any significant differences females (Mann-Whitney, U ϭ 22, P ϭ 0.29; in the sperm remaining in their palps and, as Kruskal-Wallis, df ϭ 2, H ϭ 3.1, P ϭ 0.22, aconsequence,nodifferenceinthespermre- respectively; Table 1); however, there was a leased to females (Mann-Whitney, sperm re- trend for copulation to be shortest with non- maining U ϭ 20, P ϭ 0.21; sperm released U virgin females and longest with wild-caught ϭ 25, P ϭ 0.46, respectively; Table 2). The females. Wild-caught males observed in the sperm available prior to copulation was esti- laboratory had a total copulation duration that mated from the average sperm count of palps was longer, but not significantly so, than those from virgin laboratory-reared males. Although observed in the field (Mann-Whitney, U ϭ 1, there were no differences in sperm release be- P ϭ 0.25; Table 1). tween males, there was variation in sperm re- Mate-Avoidance Behavior: In one case, the lease and palp use within a male. Most males female mate-avoidance behavior already not- released sperm from each palp and used both ed in the field was seen again in the labora- palps exactly once during mating (n ϭ 10/16). tory. During one staged mating between wild- Two males mating non-virgin females re- caught spiders, after pairing and engaging leased sperm from both palps but inserted the chelicerae, the female began resisting sperm first palp twice, after the insertion of the sec- transfer. She kept her abdomen away from the ond palp (n ϭ 2/16). Some males released male’s palps and walked backwards down the sperm from only one palp (n ϭ 4/16). These branch towards the aquarium water. Upon four males (three mated to virgin females, one contacting the water, the female released her mated to a non-virgin female) inserted the first hold on the branch and floated on the water palp only once during copulation. The sperm surface next to the branch. Yet, perhaps be- release by a male was not correlated with the cause there was no current pulling her away, overall duration of copulation (Spearman rank the male was able to keep his chelicerae in- correlation, rho ϭ 0.250, Z ϭ 0.97, P ϭ 0.33). terlocked with the female and continue court- Sperm Release From ‘‘One-insertion’’ ing. Courtship lasted one hour and 31 s, pos- Males: To look at natural variation in sperm sibly another measure of female resistance as number as well as sperm release, a subset of typical courtship duration was approximately virgin laboratory-reared males (n ϭ 3) and three minutes. The male managed to insert wild-caught males (n ϭ 3) were allowed to only his right palp for 11 min, but did not copulate with only one palp before being sac- release any sperm (as determined afterwards rificed for sperm counting. These males are by comparing the sperm remaining in the used described as ‘‘one-insertion’’ males. Unused palp to that in the unused palp), which may palps from virgin laboratory-reared males indicate that a more cryptic form of female contained significantly less sperm than unused resistance exists (this male was excluded from palps from wild-caught males (X¯ ϭ 483,000 394 THE JOURNAL OF ARACHNOLOGY

Table 3.—Summary of palp insertion, sperm release, and sperm remaining for the ﬁrst palp only for virgin males mated to virgin and non-virgin females, and wild-caught males mated to wild-caught females in the laboratory, X¯ Ϯ SE (n). *To calculate release, the sperm retained in the mated palp was subtracted from the estimate of sperm available prior to copulation.

Female type Virgin Non-Virgin Wild-Caught Palp insertion (s) 563 Ϯ 49 (8) 294 Ϯ 88 (8) 659 Ϯ 86 (3) Sperm released* 477,000 Ϯ 12,000 (8) 456,000 Ϯ 23,000 (8) 470,000 Ϯ 9,000 (3) Sperm remaining 19,000 Ϯ 9,000 (8) 27,000 Ϯ 23,000 (8) 14,000 Ϯ 9,000 (3)

¯ Ϯ 129,000, X ϭ 1,026,000 Ϯ 240,000; Mann- ternity patterns in this species. S2 is the pro- Whitney, U ϭ 9, P ϭ 0.05). This subset of portion of sperm from the second male to mate males did release fully from the palp used in (S2 ϭ Sperm2nd Male/SpermTotal). Using the con- mating, as the amount of palp sperm remain- vention of Christenson & Cohn (1988), sperm ing was similar for virgin laboratory-reared release patterns were categorized as first-male ¯ and wild-caught males (X ϭ 14,500 Ϯ 9,000, biased (S2 Յ 0.33), mixed (S2 ϭ 0.33–0.66), ¯ X ϭ 14,000 Ϯ 9,000; Mann-Whitney, U ϭ 4, and last-male biased (S2 Ն 0.66). Sperm release P ϭ 0.83). patterns were variable: mixed (n ϭ 4), last- Sperm Release from First Palp Insertion: male biased (n ϭ 3) and first-male biased (n ϭ First palp insertion duration and sperm release 1). Each case of bias, either first- or last-male, were compared for males mated to virgin, resulted from a male releasing sperm from only non-virgin and wild-caught females (Table 3). one palp whereas the other male released from The duration of the first palp insertion was both palps, possibly gaining an advantage in significantly shorter for males mated to non- sperm competition. Overall, S2 values averaged virgin females than for males mated to either 0.51 Ϯ 0.05 (n ϭ 8). virgin or wild-caught females (Kruskal-Wal- lis, df ϭ 2, H ϭ 7.5, P ϭ 0.03; Table 3). Using DISCUSSION the conservative method, the sperm retained Glenognatha emertoni are well-hidden in and the sperm released from the first palp was the daytime, and emerge at dusk to spin orb not significantly different for males mated to webs over the stream. Male G. emertoni do virgin, non-virgin and wild-caught females not cohabitate with or guard females on their (Kruskal-Wallis, sperm retained, df ϭ 2, H ϭ orb-webs. Instead, males leave females im- 1.57, P ϭ 0.46; sperm released, df ϭ 2, H ϭ mediately after mating, similar to Tetragnatha 0.56, P ϭ 0.76; Table 3). However, the cal- species (LeSar & Unzicker 1978; West & Toft culated sperm release for wild-caught males 1999; Danielson-François 2002; Danielson- would be twice as much if a less conservative François & Bukowski 2005), but unlike phol- estimate was used: taking the average wild- cid species that commonly have cohabitation caught, rather than virgin sperm count, to ap- (Eberhard & Bricenõ 1983; Eberhard et al. proximate the amount of sperm available be- 1993; Blanchong et al. 1995; Kaster & Jakob fore copulation (resulting in wild-caught male 1997). During the day, G. emertoni females sperm release of X¯ ϭ 998,000 Ϯ 9,000 per are found in retreats, but not alone—typically palp, Kruskal-Wallis, df ϭ 2, H ϭ 7.3, P ϭ more than one male is nearby. Some females 0.03; compare to Table 3). Two males did not had as many as five surrounding males, but release any sperm and so were not included whether this was due to chance, mate-guard- in these statistical analyses (one male was ing of or a male preference for particular fe- mated to a non-virgin female, the other male males remains to be tested (c.f. Danielson- was mated to a wild-caught female). François et al. 2002). Sperm Release Patterns: Icalculatedtherel- Males do not appear to discriminate be- ative sperm release by the first and second lab- tween virgin, non-virgin, and wild-caught fe- oratory-reared male for each doubly-mated females, being equally likely to court and mate male in the study. This calculation is useful to each type of female. Despite this lack of dis- provide a null hypothesis for future tests of pa- crimination, there were some differences in DANIELSON-FRANÇOIS—BEHAVIOR& SPERMRELEASEINA HAPLOGYNE 395 mating behavior. Laboratory-reared males in sperm release. Whether releasing sperm courted non-virgins significantly longer than from only one palp during mating was due to virgin females, and wild-caught adult males cryptic actions by the male or female remains tended to get to the courtship phase more to be seen. quickly than laboratory-reared virgin males. Huber & Eberhard (1997) suggest that hap- Glenognatha emertoni mating behavior was logyne males might be able to influence pa- unique in having a vibratory phase of court- ternity by repositioning rival males’ sperm ship that was absent in its congeners, but oth- due to the short insemination duct found in erwise was similar to that of other tetragna- most haplogyne taxa. In the haplogynes Phy- thines in using chelicerae as holdfasts during socyclus globosus and Leucauge mariana mating (Barrows 1919; LeSar & Unzicker (Taczanowski 1881), males had extensive pal- 1978; Huber 1998; West & Toft 1999; Dan- pal movements that differed for virgin and ielson-François 2002; Danielson-François & non-virgin females (Huber & Eberhard 1997; Bukowski 2005). After cheliceral engage- Eberhard & Huber 1998). In the haplogyne ment, latency to courtship was brief (less than Pholcus phalangioides, second males gain one min) and vibratory courtship (leg-tapping) higher paternity even though they copulate for lasted several min. Overall mating duration a shorter amount of time than first males (Yo- with laboratory-reared males was approxi- ward 1998). A detailed examination of P. mately 20 min, similar to the 15 min observed phalangioides revealed that paternity was in- for G. foxi (Barrows 1919). Wild caught- fluenced by the number of pedipalp move- males mated in the laboratory had a 40 min ments made by the second male, and Scha¨fer duration, whereas those observed in the field & Uhl (2002) suggest that these movements had a shorter duration, about 17 min. could remove rival’s sperm. Consistent with Consistent with the lack of overt discrimi- the above studies, the flexible and short com- nation between females based on mating his- mon insemination duct in G. emertoni would tory, G. emertoni males released equivalent seem to make them particularly susceptible to amounts of sperm to virgin, non-virgin, and this sort of manipulation (Danielson-François wild-caught females. This pattern of sperm re- 2002). lease was also noted in the related haplogyne The short and flexible nature of the haplo- Tetragnatha versicolor Walckenaer 1842 gyne insemination duct may also increase the (Danielson-François & Bukowski 2005), but evolutionary lability of palp insertion patterns. was distinctly different from the sperm release Huber & Senglet (1997) argued that the ab- pattern of the related entelegyne orb-weaver sence of a ‘‘lock-and-key’’ fit between male Nephila clavipes (Linnaeus 1767), whose and female genitalia in taxa within the ente- males preferentially release sperm to virgin legyne family Tetragnathidae made it possible females, but release none, or only a greatly to switch from the usual ipsi- to contra-lateral reduced amount, during mating to non-virgin insertion, inserting the right palp into the left females (Christenson & Cohn 1988). side of the epigynum. In contrast to the rest Despite a lack of significant differences in of the tetragnathine genera (i.e., Tetragnatha sperm release to virgin and non-virgin fe- and Pachygnatha) and Leucauge, G. emertoni males, there was variation among G. emertoni insertion patterns appear to be ipsilateral. Sim- males in sperm release. Calculated for each ilarly, G. heleios also has an ipsilateral palp doubly-mated female in the study, the result- insertion pattern (Edwards & Senske 2001). ing proportion of sperm released by the sec- The palp insertion patterns of G. emertoni ond male to mate resulted in a range of first- were based on behavioral observations. Fur- to last-male bias (0.33–0.68), and yielded an ther confirmation of this result would require average sperm release of 0.51 Ϯ 0.05, similar examination of pairs flash-frozen in copula to that observed for T. versicolor (Danielson- (c.f. Huber & Senglet 1997) to physically de- François & Bukowski 2005). First and second termine which side (right or left) the palp en- males typically released sperm from both ters. palps, and all males inserted both palps at Female G. emertoni appear to have a suite least once during mating, but sometimes a of resistance behaviors, one of which may male would release sperm from only one of have its origins in a startle behavior common his palps and not the other, resulting in a bias to both sexes—dropping out of the web and 396 THE JOURNAL OF ARACHNOLOGY into the water below. The floating ability of Females actively resisted courting males by spiders is influenced, in part, by the hydro- releasing from them and floating downstream, phobicity of the cuticle and the density of and may have a cryptic method of preventing hairs (Suter et al. 2004). Although the hydro- sperm release during copulation. Future stud- phobicity of cuticle in G. emertoni is un- ies will examine the relationship among sperm known, this small spider has a dense set of release, mating behavior, and paternity pat- hairs covering the body that may aid its ability terns in G. emertoni and other haplogyne spe- to float. Dropping into the current and floating cies. downstream may prove to be a common mode of transport once more species are examined, ACKNOWLEDGMENTS but currently there is less information on float- For spider collecting and field assistance I ing as a means of travel even though loco- would like to thank P. Fenberg, L. Krebs, and motion across the water surface (e.g., walking, S. Danielson-François. For permits and sup- rowing, and galloping) is a well-documented port I would like to thank Fort Huachuca phenomenon in spiders (Suter et al. 2003; Wildlife Biologist H. S. Stone. I would also Stratton et al. 2004). like to thank G. Bodner, P. Cushing, B. Cutler, Eberhard (2004) suggested that spiders are B. Huber, A. Little, W. Maddision, T. Markow, excellent organisms for the study of sexual se- D. Mott, D. Papaj, R. Smith, G. Stratton, and lection, and in particular sperm competition. an anonymous reviewer for their insightful As a general method, examining the amount comments on earlier drafts of this manuscript. of sperm released per male during mating can This material is based upon work supported be used to generate a null hypothesis for test- by the National Science Foundation grant No. ing the mechanisms underlying paternity pat- 0073266 and the University of Arizona Center terns in spiders. One such null hypothesis for Insect Science. would be that the proportion of sperm from LITERATURE CITED the second male to mate, S2, would be equivalent to—under a simplistic ‘‘lottery’’ model Andrade, M.D. 1996. Sexual selection for male sac- (Parker 1998)—the expected paternity for the rifice in the Australian redback spider. Science second male (P , sensu Boorman & Parker 271:70–72. 2 Austad, S.N. 1984. Evolution of sperm priority pat- 1976). This is a null hypothesis that can and terns in spiders. Pp. 223–249. In Sperm Com- should be explicitly tested in future paternity petition and the Evolution of Animal Mating studies. Given the pattern of sperm release Systems (R.L. Smith, ed.). Academic Press, New found in G. emertoni, the null hypothesis York. based only on sperm release would be mixed Barrows, W.M. 1919. The taxonomic position of paternity, all else being equal. The relation- Mysmena bulbifera (Glenognatha bulbifera) ship of sperm release to paternity patterns has Banks, with some observations on its habits. yet to be studied for G. emertoni, and if male Ohio Journal of Science 19:210–212. or female manipulation of sperm release and Blanchong, J.A., M.S. Summerfield, M.A. Popson storage occurs, it is unlikely that such a sim- & E.M. Jakob. 1995. Chivalry in pholcid spiders revisited. Journal of Arachnology 23:165–170. plistic scenario will hold. Future work is need- Boorman, E. & G. Parker. 1976. Sperm (ejaculate) ed to determine the actual paternity pattern by competition in Drosophila melanogaster, and the obtaining empirical P2-values and comparing reproductive value of females to males in rela- them to sperm release within the same study. tion to female age and mating status. Ecological In summary, Glenognatha emertoni mating Entomology 1:145–155. behavior was fairly simple, with a short laten- Bukowski, T.C. & T.E. Christenson. 1997. Deter- cy to copulation and a phase of vibratory minants of sperm release and storage in a spiny courtship before a single palp insertion (ipsi- orbweaving spider, Micrathena gracilis. Animal lateral) from each palp (right and left) to trans- Behaviour 53:381–395. fer sperm. Males left immediately after cop- Bukowski, T.C., C.D. Linn & T.E. Christenson. 2001. Copulation and sperm release in Gastera- ulation and did not cohabitate with females on cantha cancriformis (Araneae: Araneidae): dif- their orb webs. Males released equivalent ferential male behavior based on female mating amounts of sperm to females, regardless of history. Animal Behaviour 62:887–895. their mating history, similar to the only other Christenson, T.E. & J. Cohn. 1988. Male advantage haplogyne studied for sperm release behavior. for egg fertilization in the golden orb-weaving DANIELSON-FRANÇOIS—BEHAVIOR& SPERMRELEASEINA HAPLOGYNE 397

spider (Nephila clavipes). Journal of Compara- Griswold, C.E., J.A. Coddington, N. Platnick & R. tive Psychology 102:312–318. Forster. 1999. Towards a phylogeny of entele- Danielson-François, A.M. 2002. Variation in tetra- gyne spiders (Araneae, Entelegynae). Journal of gnathid spermathecal structures and sperm com- Arachnology 27:53–63. petition with descriptions of natural history. Ph. Hormiga, G. & H.G. Do¨bel. 1990. A new Gleno- D. dissertation, Department of Ecology and Evo- gnatha (Araneae, Tetragnathidae) from New Jer- lutionary Biology, University of Arizona, Tuc- sey, with redescriptions of G. centralis and G. son, Arizona USA. minuta. Journal of Arachnology 18:195–204. Danielson-François, A.M., C. Fetterer & P. Small- Hormiga, G., W.G. Eberhard & J.A. Coddington. wood. 2002. Body condition and mate choice in 1995. Web-construction behavior in Australian Tetragnatha elongata (Araneae: Tetragnathidae). Phonognatha and the phylogeny of Nephiline The Journal of Arachnology 30:20–30. and tetragnathid spiders (Araneae, Tetragnathi- Danielson-François, A.M. & T.C. Bukowski. 2005. dae). Australian Journal of Zoology 43:313–364. Female mating history influences copulation be- Huber, B. 1998. Spider reproductive behaviour: a havior but not sperm release in the orb-weaving review of Gerhardt’s work from 1911–1933, with spider Tetragnatha versicolor (Araneae, Tetra- implications for sexual selection. Bulletin of the gnathidae). Journal of Insect Behavior 18:131– British Arachnological Society 11:81–91. 148. Huber, B. & W.G. Eberhard. 1997. Courtship, cop- Eberhard, W.G. 1996. Female Control: Sexual Se- ulation and genital mechanics in Physocyclus lection by Cryptic Female Choice. Princeton globosus (Araneae, Pholcidae). Canadian Journal University Press, Princeton, New Jersey. 472 pp. of Zoology 74:905–918. Eberhard, W.G. 2004. Why study spider sex: special Huber, B. & A. Senglet. 1997. Copulation with con- traits of spiders facilitate studies of sperm com- tralateral insertion in entelegyne spiders (Ara- petition and cryptic female choice. The Journal neae: Entelegynae: Tetragnathidae). Netherlands of Arachnology 32:545–556. Journal of Zoology 47:99–102. Eberhard, W.G. & R.D. Bricenõ. 1983. Chivalry in Kaster, J.L. & E.M. Jakob. 1997. Last-male sperm pholcid spiders. Behavioral Ecology and Socio- priority in a haplogyne spider (Araneae, Pholci- biology 13:189–195. dae): correlations between female morphology Eberhard, W.G., S. Guzman-Gomez & K. Catley. and patterns of sperm usage. Annals of the En- 1993. Correlation between spermathecal mor- tomological Society of America 90:254–259. phology and mating systems in spiders. Biolog- LeSar, C. & J. Unzicker. 1978. Life history, habits ical Journal of the Linnean Society 50:197–209. and prey preferences of Tetragnatha laboriosa Eberhard, W.G. & B. Huber. 1998. Courtship, cop- (Araneae: Tetragnathidae). Environmental Ento- ulation, and sperm transfer in Leucauge mariana mology 7:879–884. (Araneae, Tetragnathidae) with implications for Levi, H.W. 1980. The orb-weaver genus Mecyno- higher classification. Journal of Arachnology 26: gea, the subfamily Metinae and the genera Pach- 342–368. ygnatha, Glenognatha and Azilia of the Subfam- Edwards, R.L. & C.W. Senske. 2001. Observations ily Tetragnathidae North of Mexico (Araneae: on the breeding behavior of the spider Gleno- Araneidae). Bulletin of the Museum of Compar- gnatha heleios (Araneae: Tetragnathidae). Ento- ative Zoology 149:175. mological News 112:71–72. Parker, G. 1998. Sperm competition and the evo- Elgar, M.A. 1998. Sperm competition and sexual lution of ejaculates: towards a theory base. Pp. selection in spiders and other arachnids. Pp. 307– 3–54. In Sperm Competition and Sexual Selec- 339. In Sperm Competition and Sexual Selection tion (T.R. Birkhead & A.P. Moller, eds.). Aca- (T.R. Birkhead & A.P. Moller, eds.). Academic demic Press, New York. Press, New York. Platnick, N.I. 2006. The World Spider Catalog. Ver- Gerhardt, U. 1921. Vergleichende Studien u¨ber die sion 5.5. American Museum of Natural History, Morphologie des ma¨nnlichen Tasters und die New York. Online at: http://research.amnh.org/ Biologie der Kopulation der Spinnen. Archiv fu¨r entomology/spiders/catalog/index.html. Naturgeschichte 87:78–245. Scha¨fer, M.A. & G. Uhl. 2002. Determinants of pa- Gerhardt, U. 1927. Neue biologische Untersuchun- ternity success in the spider Pholcus phalangioi- gen an einheimischen und ausla¨ndischen Spin- des (Pholcidae: Araneae): the role of male and nen. Zeitschrift fu¨r Morphologie und O¨ kologie female mating behaviour. Behavioral Ecology der Tiere 8:96–186. and Sociobiology 51:368–377. Griswold, C.E., J.A. Coddington, G. Hormiga & N. Simmons, L.W. & M.T. Siva-Jothy. 1998. Sperm Scharff. 1998. Phylogeny of the orb-web build- competition in insects: mechanisms and the po- ing spiders (Araneae, Orbiculariae: Deinopoidae, tential for selection. Pp. 341–434. In Sperm Araneoidea). Zoological Journal of the Linnean Competition and Sexual Selection (T.R. Birkhead Society 123:1–99. & A.P. Moller, eds.). Academic Press, New York. 398 THE JOURNAL OF ARACHNOLOGY

Simon, E. 1895. Histoire naturelle des Araigne´es. Tetragnatha extensa (Araneae: Tetragnathidae). Deuxie`me e´dition. Tome premier. Quatrie`me fas- Journal of Insect Behavior 12:433–450. cicule. Librarie Encyclope´dique de Roret, Paris. Wiehle, H. 1967. Meta,-eine semientelegyne Gat- 761–1084 pp. tung der Araneae (Arach.). Senckenbergiana Stratton, G.E., R.B. Suter & P.R. Miller. 2004. Evo- Biologica 48:183–196. lution of water surface locomotion by spiders: a Yoward, P. 1996. Spider sperm competition: the comparative approach. Biological Journal of the conduit/cul-de-sac hypothesis—a route to under- Linnean Society 81:63–78. standing or a dead end? D. Phil. Thesis, York Suter, R.B., G.E. Stratton & P.R. Miller. 2003. Wa- University, UK. ter surface locomotion by spiders: distinct gaits Yoward, P. 1998. Sperm competition in Pholcus in diverse families. The Journal of Arachnology phalangoides (Fuesslin, 1775) (Araneae, Pholci- 31:428–432. dae)—Shorter second copulations gain a higher Suter, R.B., G.E. Stratton & P.R. Miller. 2004. Tax- paternity reward than ﬁrst copulations. Pp. 167– onomic variation among spiders in the ability of 170. In Proceedings of the 17th European Col- repel water: surface adhesion and hair density. loquium of Arachnology, Edinburgh, 1997. (P.A. The Journal of Arachnology 32:11–21. Selden, ed.). British Arachnological Society, Walker, W.F. 1980. Sperm utilization strategies in non- Burnham Beeches, UK. social insects. American Naturalist 115:780–799. West, H.P. & S. Toft. 1999. Last male sperm priority Manuscript received 21 July 2004, revised 28 Au- and the mating system of the haplogyne spider gust 2006.