Sperm Storage in Spermathecae of the Great Lamper Eel, Amphiuma Tridacfyhm (Caudata: Amp Hi U Midae)

Home , Amphiuma, Amphiuma tridactylum, Axolotl

JOURNAL OF MORPHOLOGY 230:79-97 (1996)

Sperm Storage in Spermathecae of the Great Lamper Eel, Amphiuma tridacfyhm (Caudata: Amp hi u midae)

DAVID M. SEVER, J. SEAN DOODY, COURTNEY A. REDDISH, MICHELLE M. WENNER, AND DON R. CHURCH Department of Biology, Saint Mary's College, Notre Dame, Indiana 46556 (D.M.S., C.A.R., M.M. W.); Department of Biological Sciences, Southeastern Louisiana University, Hammond, Louisiana 70402 (J.S.D., D.R.C.)

ABSTRACT The spermathecae of ten female Amphiuma tridactylum were examined by light and electron microscopy during the presumed mating and ovipository seasons (March-August) in Louisiana. Spermathecae were simple tubuloalveolar glands in the dorsal wall of the cloaca. Six of the ten specimens were vitellogenic, and all of these specimens contained sperm in their spermathecae and had secretory activity in the spermathecal epithelium. Two nonvitellogenic females also had sperm in their spermathecae and active epithelial cells, whereas the other nonvitellogenic females lacked stored sperm and secretory activity in the spermathecae. In specimens storing sperm from March-May, the sperm were normal in cytology, and secretory vacuoles were contained within the epithelium. In the August sample, however, evidence of sperm degradation was present, and secretory material had been released into the lumen by an apocrine process. We therefore hypothesize that the spermathecal secretions function in sperm degeneration. Q 1996 Wiley-Liss, Inc.

The Amphiumidae consists of three spe- of A. tridactylum and described these glands cies of Amphiuma, elongate, fossorial sala- in A. means and A. phloleter. In addition, manders with reduced limbs and one exter- Sever ('gla, '94) reported on the phylogeny nal gill slit, a paedomorphic character of the spermathecae and other cloacal glands (Duellman and Trueb, '86). Two of the spe- in the Amphiumidae. cies are aquatic and often exceed 70 cm snout- Kreeger ('42) stated that female A. tridac- vent length (SVL); A. means occurs in the tylum could store sperm in their spermathe- coastal plain from Virginia to Louisiana, and cae for 7 or 8 months, and Baker ('62) found A. tridactylum occurs from western Alabama that isolated females could retain living sperm to Texas and up the Mississippi River valley for 15 months. In females of other sala- to western Kentucky. A. phloleter is a dwarf manders, the longest period that Sever ('95) (20-25 cm SVL), terrestrial species limited to felt was reliable for duration of sperm stor- the Florida panhandle and southern Ala- age was 6 months. The best way to ascertain bama (Conant and Collins, '91). In the Missis- the cytological condition of sperm during stor- sippi River valley, A. tridactylum is com- age is by use of electron microscopy, but no monly called the "great lamper eel" (Parker, ultrastructural studies of the spermathecae '37; Baker, '45; Dundee and Rossman, '89). of Amphiuma have heretofore existed. We A number of reports exist on the reproduc- recently had the opportunity to cytologically tive biology of Amphiuma, but most of these examine the spermathecae of ten females are anecdotal and some are contradictory sacrificed during the mating and ovipository (Table 1). The majority of the literature con- period, and herein we present our findings on cerns A. tridactylum. Sperm in the female the ultrastructure of sperm storage in this cloaca was first observed by Davison (1895). species. These results are then compared with The only report of mating behavior is an those on the ultrastructure of the sperma- observation of copulation by cloacal apposi- thecae of other salamanders (for recent re- tion (Baker et al., '47). Sperm storage glands, views see Sever, '95; Sever et al., '96). spermathecae, in the female cloaca were first described by Kreeger ('42), and Sever ('92a) Address reprint requests to David M. Sever, Department of extended her observations on spermathecae Biology, Saint Mary's College, Notre Dame, IN 46556. o 1996 WILEY-LISS, INC. 80 D.M. SEVER ET AL MATERIALS AND METHODS light microscopy (LM) or for embedding in Great lamper eels were collected from road- epoxy resin for thin (LM) or ultrathin sec- side ditches at three sites, each within 10 km tions for transmission electron microscopy of Hammond, Tangipahoa Parish, Louisiana. (TEM). Testes and vasa deferentia were re- We obtained specimens monthly from Febru- moved from males, and cloacae were dis- sected from females. For all specimens, tis- ary-May 1995. We captured most individuals sues were initially fixed in 10% neutral by hand but also used a dipnet, and a few individuals were caught in funnel traps baited buffered formalin (NBF) (for LM) or in a 1:l with chicken necks. We concentrated collect- solution of 2.5% glutaraldehyde in Millonig's phosphate buffer at pH 7.4 and 3.7% formal- ing efforts at night after rain, when lamper dehyde buffered to pH 7.2 with monobasic eels were more active; many were obtained and dibasic phosphate (TEM). Carcasses of from ditches that were dry prior to rainfall. all specimens are stored in NBF in the re- Because females become scarce in sum- search collections at Saint Mary's College. mer, presumably as a result of nesting (Cagle, For paraffin infiltration prior to sectioning '48; Fontenot, '901, we placed 22 presumed for LM, the tissue was rinsed in water after adult females captured between 9 and 13 fixation, dehydrated in ethanol, cleared in May into two outdoor cattle tanks (2.0 m Histosol (National Diagnostics, Inc., Man- diameter, 0.7 m sidewall height) in Hammond, ville, NJ), and embedded in paraffin. Sections Louisiana. Mature females were considered (10 km) were cut with a rotary microtome, those >33 cm total length (Fontenot, 'go), and a&ed to albuminized slides, and alter- and females were identified by black pigmen- nate slides were stained with hematoxylin- tation in the cloaca1 wall (Baker, '37). Cattle eosin (HE) (for general cytology) or alcian tanks were filled with water to a level of blue 8GX (AB) (for primarily carboxylated approximately 15cm and contained mud from glycosaminoglycans) at pH 2.5 followed by a ditch at one of the collecting sites. Eight the periodic acid-Schiff method (PAS) (neu- plastic nesting boxes (25 x 50 x 20 cm) con- tral carbohydrates and sialic acids). Proce- taining mud were placed at staggered heights dures followed Kiernan ('90). in each tank, and some individuals entered After initial fixation, tissues prepared for the boxes. Plywood sheets were placed over plastic infiltration prior to sectioning for LM the cattle tanks to prevent further filling due and TEM were trimmed into 1.5 mm blocks, to rainfall. Tanks were examined for eggs rinsed in Millonig's buffer, postfixed in 2% every few days until 25 July when other osmium tetroxide, dehydrated in ethanol, circumstances dictated that the observations cleared in propylene oxide, and embedded in must end. an epoxy resin (EMBED-812; Electron Mi- Table 2 presents the data on the specimens croscopy Sciences, Fort Washington, PA). from which spermathecae were examined by Semithin sections (0.5-1 km) for LM were light and electron microscopy. Those lulled cut with glass knives, placed on microscope on 22 August were specimens that had been slides, and stained with toluidine blue. Ultra- maintained in the cattle tanks and subse- thin sections (70 nm) for TEM were collected quently kept in a living stream (180 cm x 50 on uncoated copper grids and stained with cm x 60 cm; Frigid Units, Inc., Toledo, Ohio) solutions of uranyl acetate and lead citrate. at Saint Mary's College until time of sacri- These sections were cut with RMC XLlOOO fice. The number of yolky ovarian oocytes and RMC MT7 ultramicrotomes, and thin was counted in each female, and 11 such sections were viewed with a Hitachi H-300 oocytes from each female were measured to transmission electron microscope. Terminol- the nearest 0.1 mm with a dial caliper under ogy for sperm ultrastructure follows Picheral 5 x magnification. In addition to the females, ('79). single males were respectively collected (sacrificed) on 19 February (3 March), 6 March RESULTS (14 March), 21 April (5 May), and 9-13 May Specimens maintained in cattle tanks (18 October). During the time lamper eels were kept in Specimens were killed by immersion in cattle tanks, seven individuals died. One of 10% MS-222, and snout-vent length (SVL) the specimens retrieved from the cattle tanks was measured from the tip of the snout to the later proved to be a male (and was sacrificed posterior end of the vent. Tissues were ex- on 18 October so the testes and vasa deferen- cised from freshly killed specimens and fixed tia could be prepared for histological examina- for preparation by paraffin infiltration for tion). None of the females kept in the cattle TABLE 1. Literature on the reproductive biology ofAmphiuma Species State Spermiogenesis Vitellogenesis Mating Sperm storage Oviposition Nesting Hatching Reference Not specified October Midsummer McGregor (1889) Winter Sturdivant ('49) A. means AL Nov Ultsch and Arce-

neaiix...... ~, ('88) . NC July Brimley ('39) FL July Brimley ('10) FL January-February Weber ('44) FL October-May 15 months Baker ('62) sc July Hildebrand ('10) A. tridactylum AK August-September Hay (1888) LA August-March NovemberJuly February-April July-September Wilson ('40) LA January-March Wilson ('41) LA March-December Kreeger ('42) LA January-May MayJune Nov Cagle ('48) LA October-May 15 months April-May Fall Baker ('62) LA December-February November-April February-winter Rose ('67) LA September-May June June-November November- Fontenot ('90) December TN May August- Davison (1895) September TN August-winter Parker ('37) TN July August August-September Baker ('37) TN July Baker et al. ('47) 82 D.M. SEVER ET AL. TABLE 2. SDecimens used in this study' mathecae, indicative of recent mating. AS Ovarian oocytes noted by Sever ('92a),the spermathecae were numerous, simple, tubuloalveolar exocrine Mean glands opening into the roof of the cloaca. Collected Sacrificed SVL Sperm2 N diameter SD Sperm were not present in all spermathecae 19Feb 3Mar 55.0 + 240 3.7 0.21 observed in a given section and seemed most 6Mar 14 Mar 65.0 + 269 4.1 0.43 abundant in the widened lumina of the alveo- 6Mar 14Mar 59.6 + - l0Ap 21Ap 53.0 0 - lar end of the glands (Fig. 2A). The epithelial 10Ap 24Ap 40.4 + 93 4.0 0.28 cells were cuboidal to columnar, and the 21 Ap 5May 56.2 0 - largely euchromatic nuclei were basal and 21 Ap 5 May 48.8 + 152 4.7 0.21 oriented to the long axes of the cells. Round 9-13May 22Aug 59.0 0 137 5.8 0.42 9-13May 22Aug 51.0 + - to oval secretory vacuoles (0.3-0.7 km) were 9-13May 22Aug 68.8 + 254 6.0 0.46 abundant in the apical cytoplasm, and intercellular canaliculi were narrow (Fig. 2B). 'Measurements are in millimeters. 'Sperm present ( +) or sperm absent ( - 1 Abundant vesicles, ribosomal complexes, microfilaments, and elongate mitochondria with tubular cristae were associated with secre- tanks oviposited eggs while in the tanks or tory vacuoles in the perinuclear cytoplasm when subsequently maintained in a living (Fig. 2C). stream. Mature secretory vacuoles usually contained a circular, electron-dense area (0.2- Male reproductive cycle 0.3 km) along one portion of the circumfer- Since the male spermatogenic cycle has ence and a less electron-dense (more been described elsewhere in detail (Wilson, flocculent) area that was often larger and '40; Baker, '62; Rose, '67), we examined only more ovate (Fig. 20. Immature condensing males sacrificed in March, May, and October vacuoles were composed entirely of the floccu- to verify the pattern. We found sperm in the lent material, so the electron-dense portion vasa deferentia of all males, and sperm were may result from compaction of the flocculent relatively most numerous in vasa deferentia material. Since alpha and beta particles of of specimens sacrificed in March. The testes glycogen were absent, the secretory vacuoles of the specimen examined from October con- were responsible for the PAS+ reaction in tained lobules filled with spermatids and sper- paraffin sections. The combination of ribo- matozoa, while testes of the specimens from somal organelles responsible for peptide syn- March contained few sperm and those of the thesis and a PAS+ reaction indicates that specimen sacrificed in May were totally evacu- the product contained glycoproteins (Kiernan, ated of sperm. '90). No evidence of release of the secretory product was found in specimens sacrificed Female reproductive cycle March-May. Data on the presence of sperm and vitello- Portions of sperm cells were sometimes genesis (presence of enlarged yolky ovarian found embedded in the apical cytoplasm of oocytes) are given in Table 2 for females the spermathecal epithelium (Fig. 3A), but sacrificed between 3 March and 22 August. most sperm were observed in the lumen in Females containing vitellogenic follicles pos- small clusters in which sperm had a similar sessed sperm in their spermathecae, and oo- orientation (Fig. 3B,C). These small clusters cytes in specimens sacrificed March-May were often adjacent to other groups of sperm were smaller than in those sacrificed in Au- with different orientations. The sperm ap- gust (Table 2; Fig. 1A,B). Four females lacked peared normal in cytology (Fig. 3B-D). Mito- vitellogenic follicles. Two of these females, chondria were numerous around the axial one sacrificed 14 March (Fig. 1C) and the rod of the principal piece of the tail and were other sacrificed 22 August (Fig. lD), contained characterized by lamellar cristae (Fig. 3D). sperm in their spermathecae (Table 2). Gravid females sacrificed in August Gravid females sacrificed March-May Two females (59.0 and 68.8 cm SVL) col- Four females (40.4-65 cm SVL) collected lected between 9 and 13 May, maintained in a 19 February-21 April and sacrificed 3 cattle-tank until 25 July, and subsequently March-5 May possessed 93-269 yolky ovar- kept in a living stream until sacrificed on 22 ian oocytes 3.7-4.7 mm mean diameter and August contained 137-254 yolky ovarian fol- contained abundant sperm in their sper- licles 5.8-6.0 mm mean diameter and had SPERM STORAGE IN AMPHIUMA 83

Fig. 1. Female Amphiumu triductylum showing rela- C: Specimen 59.6 mm SVL collected 6 March and sacri- tive development of oocytes in specimens in which sper- ficed 14 March lacking enlarged yolky oocytes. D: Speci- mathecae contained sperm. A: Specimen 40.4 mm SVL men 51.0 mm SVL collected between 9 and 13 May and collected 10 April and sacrificed 24 April with 93 yolky sacrificed 22 August lacking enlarged yolky oocytes. Fb, oocytes 4.0 mm mean diameter. B: Specimen 59.0 mm fat bodies; In, intestines; Lg, lung; 00,enlarged yolky SVL collected between 9 and 13 May and sacrificed 22 oocytes; Ov, ovary; So, stomach; Ub,urinary bladder. August with 137 yolky oocytes 5.8 mm mean diameter. 84 D.M. SEVER ET AL

Fig. 2. FemaleAmphiuma tridactylum, 40.4 mm SVL. graph showing detail of cytoplasm containing synthetic Specimen collected 10 April and sacrificed 24 April with organelles and the relative density of secretory vacuoles. 93 yolky ovarian oocytes 4.0 mm mean diameter. A Cr, cristae; Cv, condensing vacuole; Dm, dense material; Light micrograph of plastic semithin section stained with Fm, flocculent material; Ic, intercellular canaliculi; Lu, toluidine blue showing overview of spermathecal tubules lumen; Mf, microfilaments; Mi, mitochondria; Nu, epithe- containing sperm and surrounding connective tissue. B: lial cell nucleus; Po, polyribosomes; Se, spermathecal Electron micrograph showing supranuclear cytoplasm epithelium; Spl, sperm in the lumen; Sv, secretory vacu- with numerous secretory vacuoles. C: Electron micro- oles; Tp, tunica propria; Ve, vesicles. Fig. 3. Female Amphiuma tridactylum, electron mi- primarily principal pieces of the tail in adjacent groups of crographs. A Specimen 48.8 mm SVL collected 21 April sperm. D: Shows detail of the middle piece of the tail, and and sacrificed 5 May with 152 yolky ovarian oocytes 4.7 the inset depicts the lamellar cristae characteristic of mm mean diameter, showing sperm in contact with the mitochondria associated with the middle piece of the tail. apical spermathecal epithelium. B-D: Specimen 40.4 Ac, apical cytoplasm of the spermathecal epithelium; Af, mm SVL collected 10 April and sacrificed 24April with 93 axial fiber; Ax,axonemal complex; Cr, cristae; Lu, lumen; yolky ovarian oocytes 4.0 mm mean diameter, showing Mi, mitochondria; Mpt, middle piece of the tail; Mt, portions of sperm cells in lumen of a spermathecal tu- marginal filament; Ppt, principal piece of the tail; Sn, bule. Clusters of sperm have similar orientations. B: sperm nucleus; Um, undulating membrane. Shows mostly middle pieces of the tail. C: Illustrates 86 D.M. SEVER ET AL abundant sperm in their spermathecae. The chondrial sheaths of the sperm were dis- apical epithelial cytoplasm contained numer- rupted (Fig. 60. ous secretory vacuoles (Fig. 4A). Golgi bodies Release of secretory product into the lu- and ribosomal complexes, however, were no men was apparent in the nongravid female longer observed in the perinuclear cytoplasm sacrificed 22 August, as in gravid specimens (Fig. 4B). Secretory material was found in sacrificed at the same time. Some cytological association with sperm in the lumen (Fig. details associated with release of the secre- 4C,D). This secretory material resembled the tory product, however, were especially appar- flocculent portion of the secretory vacuoles ent in sections from the nongravid female (Fig. 4C,D). Some sperm in contact with the (Fig. 7). The release of secretory material epithelium seemed to be undergoing degrada- apparently involved the loss of apical cyto- tion as evidenced by loss of mitochondria plasm, making the mode apocrine (Fig. 7). around the axial rod (Fig. 1C). In the lumen, Clusters of secretory granules, either free in mitochondria around the axial fiber were in- the lumen or still surrounded by cell mem- vested with secretory material, and the axone- branes, were observed in lumina (Fig. 7A). ma1 complexes of sperm surrounded by secre- The flocculent portion of the secretory vacu- tory material appeared abnormal (Fig. 4D). oles dissociated from the more electron- dense portion during apocrine release of the Nongravid females containing sperm material (Fig. 7A,B). Apical areas of the sper- One female (59.6 cm SVL) collected 6 mathecal cytoplasm containing dissociated March and sacrificed 14 March and another secretory vacuoles were separated from the (51.0 cm SVL) collected between 9 and 13 nuclear regions of the cells by intracellular May and sacrificed 22 August contained nu- canaliculi that served as cleavage lines for merous sperm in their spermathecae but did the apocrine release of these portions of the not possess large yolky oocytes. The stored cytoplasm (Fig. 7B). sperm and the spermathecal epithelium were similar in cytology to those of gravid speci- Nongravid females lacking sperm in the mens sacrificed during the same periods (Figs. spermathecae 2-6). In the nongravid specimen collected 6 Two females (53.0 and 56.2 cm SVL) sacri- March and sacrificed 14 March, secretory ficed 21 April and 5 May lacked both large vacuoles were numerous, and Golgi profiles yolky ovarian oocytes and sperm in the sper- and mitochondria were abundant in the peri- mathecae (Fig. 8A). Spermathecae of these nuclear cytoplasm (Fig. 5A,B). Some sperm specimens possessed heterochromatic nuclei were embedded in the cytoplasm (Fig. 5A,D), and scant cytoplasm (Fig. 8B), and electron- but sperm in the lumen and adjacent to the dense secretory vacuoles were absent from luminaliepithelial border were normal in cy- the apical cytoplasm. The most abundant organ- tology (Fig. 5A,C). elles in the cytoplasm were mitochondria, al- In the nongravid specimen containing though scattered vesicles and areas rich in sperm that was collected 5 May and sacrificed microfilaments also were numerous (Fig. 8C). 22 August, sperm were still numerous in some luminal areas (Fig. 6A), but degrada- DISCUSSION tion of luminal sperm was evidenced by dis- Reproductive cycles ruption of the mitochondria1 sheath of the A summary of the chronology of reproduc- axial rod (Fig. 6B). Secretory vacuoles were tive events in Amphiuma based upon the abundant in the apical cytoplasm of the sper- literature is presented in Table 1. In Louisi- mathecal epithelium (Fig. 6A), but organ- ana, mating in A. tridactylum generally has elles (Golgi bodies, ribosomal complexes) in- been restricted to January-May, with most volved in further synthetic activity were not activity January-March when the male cloa- observed. In several instances, sperm were cal glands reach their maximal development observed embedded in epithelial cells, and (Wilson, '41; Cagle, '48). In Tennessee, how- these cells themselves sometimes appeared ever, mating has been ascribed to the sum- desquamated into the lumen. Both the epithe- mer months, and, indeed, the only actual lial cells, as indicated by vacuolated spaces, observation of presumed copulation in and the sperm contained within the epithe- Amphiuma was during the last 10 days of lium appeared to be degenerating (Fig. 6C). July (Baker et al., '47). The axial rods of the sperm were distorted Oviposition occurs in summer (Table I), and had lost electron density, and the mito- but nests on land have been found from SPERM STORAGE IN AMPHIUMA 87

Fig. 4. FemaleAmphzuma tridactyhm, 68.8 mm SVL, neme; Ac, apical cytoplasm of the spermathecal epithe- electron micrographs. Specimen collected between 9 and lium; Af, axial fiber; Ax,axonemal complex; Lu, lumen; 13 May and sacrificed 22 August with 254 yolky ovarian Me, myoepithelial cell nucleus; Mf, microfilaments; Mi, follicles 6.0 mm mean diameter. A Overview of sper- mitochondria; Mpt, middle piece of the tail; Mt, marginal mathecal epithelium. B: Detail of spermathecal cyto- filament; Nu, epithelial cell nucleus; Sm, secretory mate- plasm. C:Sperm and secretory material in the lumen. D rial; Spl, sperm in the lumen; Sv, secretory vacuoles; Um, Detail of principal piece of the tail of sperm in the lumen undulating membrane; Vs, vacuolated spaces. in contact with secretory material. Aax, abnormal axo- 88 D.M. SEVER ET AL.

Fig. 5. FemaleAmphiuma tridactylum, 59.6 mm SVL, cent to the spermathecal epithelium. D: Sperm nucleus electron micrographs. Specimen collected 6 March and embedded in the spermathecal epithelium. Ev, endocytic sacrificed 14 March lacking enlarged yolky oocytes. A vacuole; Go, Golgi bodies; Ic, intercellular canaliculi; Lu, Overview of spermathecal epithelium and adjacent lumi- lumen; Mi, mitochondria; Mpt, middle piece of the tail; nal areas. Areas similar to those shown in detail in B-D Mv, microvilli; Nu, epithelial cell nucleus; Ppt, principal indicated by arrowheads. B: Supranuclear cytoplasm of piece of the tail; Sn, sperm nucleus; Sv, secretory vacu- the spermathecal epithelium. C: Sperm in lumen adja- oles. Fig. 6. FemaleAmphiuma triductylum, 51.0 mm SVL, axial fiber; Ax, axoneme; Dmi, disrupted mitochondria electron micrographs. Specimen collected between 9 and from axial fiber; Lu, lumen; Mi, mitochondria; Mt, mar- 13 May and sacrificed 22 August lacking enlarged yolky ginal filament; Nu, epithelial cell nucleus; Se, spermathe- oocytes. A: Apical spermathecal epithelium and adjacent cal epithelium; Spl, sperm in the lumen; Sv, secretory lumen containing sperm. B: Detail of the middle piece of vacuoles; Um, undulating membrane; Vs, vacuolated a sperm cell in the lumen. C: Degradation of sperm in spaces. spermathecal epithelium. Aaf, abnormal axial fiber; Af, 90 D.M. SEVER ET AL.

Figure I SPERM STORAGE IN AMPHIUMA 91 June-February (Table 1).Baker ('45), how- can be obtained from the testes October- ever, notes that "every record of nests avail- April and from the sperm ducts November- able indicates that water had previously cov- May; in addition, mature sperm can usually ered the area where nests were found. It is be found in the cloaca of a female at any time. therefore believed that Amphiumae burrow Kreeger ('42) found that female A. tridacty- into the mud or crawl under a log or other lum collected and isolated in March and April protective object in the water shortly before still possessed active sperm in their cloacae in depositing eggs. . . . The water may recede April and December, respectively. Baker ('62) due to the lack of rain in the late summer or observed sperm in a female's cloaca 15 from drainage control and the nests and months after complete isolation from males. Amphiuma are . . . left in a dry area of a Such females never lay eggs in captivity, and, swamp.. . . The abundance of Amphiumae if eggs are present, they are probably re- would indicate that many nests are formed sorbed (Baker, '62). Pituitary injections gen- each year but are normally hidden under logs erally have been ineffective in inducing ovipo- or in the mud and remain below the water sition in Amphiuma (Baker, '37; Kammeraad, level." Baker ('45) believed that water may '42)' although one female A. tridactylum was be necessary to initiate hatching (as in the induced to lay 49 eggs following intramuscu- ambystomatid Ambystoma opacum). Thus, lar injection of bovine pituitary (Kammeraad, in the absence of fall rains, eggs in nests '42). Baker ('45) stated that 31% of those stranded on land may carry over from the eggs were fertile, but this information is not normal hatching time in November into later in the original paper (Kammeraad, '42). Fi- in the winter wet season in the southeastern nally, we should note Dundee and Rossman's United States, perhaps accounting for the ('89) report that Percy Viosca, an accurate reports of nests in midwinter (Parker, '37; chronicler of natural history, wrote in his Weber, '44). field notes that a female A. tridactylum col- Due to the long period of vitellogenesis lected April-May 1937 contained "two quarts (September-May) and nesting period (June- November), A. tridactylum probably is un- of eggs with active embryos," which implies able to develop the follicles necessary for ovoviviparity . annual reproduction in Louisiana (Fontenot, In the present study, males were sacrificed '90). Only 35-48% of adult females examined from October-May, and we found sperm in by various authors were gravid in any given the vasa deferentia from males in each year, leading to predictions of biennial (Wil- sample. The testes are filled with mature and son, '42; Cagle, '48) or even triennial (Fon- maturing sperm in October and are largely tenot, '90) reproduction in females. Males, evacuated by March, as found by Baker ('62). however, breed every year (Wilson, '42). Four of the ten females used were not gravid; Despite the restriction of mating to the these females were considered mature since spring (Louisiana) or summer (Tennessee), they were larger in SVL than some gravid the capability for mating and fertilization females (Table 2). Even though our sample is exists over a longer period. For A. means small, the observation that 40% of our fe- from Florida and A. tridactylum from Louisi- males were not gravid lends support to the ana, Baker ('62) noted that mature sperm notion of a biennial or triennial reproductive cycle (Fontenot, '90). The number of large, yolky oocytes in gravid females was 93-269, with the smallest oocytes in the female sacrificed 3 March (3.7 Fig. 7. Female Amphiuma triductylum, 51.0mm SVL, mm mean diameter) and the largest in fe- electron micrographs. Specimen collected between 9 and males kept in the cattle tanks and sacrificed 13 May and sacrificed 22 August lacking enlarged yolky oocytes, showing apocrine mode of secretion. A Apical on 22 August (5.8-6.0 mm mean diameter). cytoplasm and adjacent lumen showing release of secre- All of the gravid females possessed sperm in tory vacuoles and secretory material. B: Evagination of their spermathecae. These results are consis- apical cytoplasm of the spermathecal epithelium into the tent with reports of Wilson ('40) and Rose lumen. Acl, apical cytoplasm containing secretory vacuoles; C1, cleavage line; De, desmosome along an intercellu- ('67) that vitellogenesis begins in late fall and lar canaliculus; Dm, dense material; Dml, dense material winter and with findings by Wilson ('40, '4l), free in the lumen; Eac, evagination of apical cytoplasm; Cagle ('48), and Rose ('67) that mating oc- Fm, flocculent material; Fml, flocculent material free in curs winter and spring. the lumen; Lu, lumen; Mf, microfilaments; Mv, microvilli; Nu, epithelial cell nucleus; Sv, secretory vacuoles; Two of the nongravid females contained Tj, tight junction. sperm in their spermathecae. Either these 92 D.M. SEVER ET AL.

Fig. 8. Female Amphiurna tridactylum, 53.0 mm SVL. of perinuclear cytoplasm of spermathecal epithelial cells. Specimen collected 10 April and sacrificed 21 April lack- C: Electron micrograph showing detail of the cytoplasm ing enlarged yolky ovarian oocytes and possessing inac- of a spermathecal epithelial cell. Cr, cristae; Cv, condens- tive spermathecae lacking sperm. A: Light micrograph of ing vacuole; Ic, intercellular canaliculi; Lu, lumen; Mf, plastic semithin section stained in toluidine blue showing microfilaments; Mi, mitochondria; Nu, epithelial cell overview uf spermathecal tubules and surrounding con- ilildCLib, Zer. rough endoplasmic reticulum; Se, sper- nective tissue. B: Electron micrograph showing overview mathecal epithelium; Tp, tunica propria; Va, vacuoles. SPERM STORAGE IN AMPHZUMA 93 individuals stored sperm from a previous manders can store functional sperm in their breeding season when they were gravid, or spermathecae. Claims that salamanders store mating had occurred in the current breeding viable sperm for several years or between season even though the females were not successive breeding seasons are largely anec- vitellogenic. These alternatives cannot be re- dotal (Jordan, 1893; Baylis, '39; Adams, '40; solved based upon available data. However, Pool and Hoage, '73; Houck and Schwenk, no differences were observed in the cytology '84; Massey, '90). On the other hand, much of the spermathecae and stored sperm be- evidence exists for relatively short periods of tween gravid and nongravid individuals. At sperm storage in certain salamanders. Euro- least some of the sperm in the gravid and pean newts (Triturus)start to lay eggs within nongravid females sacrificed 22 August were a few days of mating (Halliday and Verrell, degenerating, and further production of secre- '84; Pecio, '92), and Wilbur ('77) reported tory vacuoles was not occurring. Sperm stor- that four species of Ambystoma (A. laterale, age between breeding seasons has not been A. maculatum, A. tigrinum, and A. trem- found in any salamander in which the annual blayi) lay their eggs within 2 days after mat- cycle of sperm storage has been studied (Sever ing. Under laboratory conditions, the axolotl et al., '96). (A. mexicanum) starts laying eggs within a Sever et al. ('96) observed hypertrophy of few hours after mating (Armstrong and the spermathecal epithelium in a nonvitello- Duhon, '89), and sperm can survive no more genic female Notophthalmus viridescens dur- than 2 weeks in the spermathecae of the ing the breeding season similar to the extent axolotl (Humphrey, '77). In studies contain- found in gravid females. The coupling among ing critical analyses of the annual cycle of vitellogenesis, hypertrophy of secondary sperm storage, the longest periods are 6 sexual characters (including the sperma- months for the salamandrid Salamandrina thecae), and mating behavior in salamanders terdigitata (Brizzi et al., '95) and the proteid needs further study (Houck and Woodley, Necturus beyeri (Sever and Bart, in press) '95). and 3 or 4 months in plethodontids such as Desmognathus "fuscus" from Louisiana Comparative cytology of the spermathecae (Marynick, '71) and Eurycea quadridigitata The seven families of salamanders compris- from Alabama (Trauth, '83). Sever ('95) pro- ing the Salamandroidea are unique among posed that short-term sperm storage (2 days vertebrates in the possession of a distinct set or less) as reported for some Ambystoma of male cloacal glands that make spermato- (Wilbur, '77; Armstrong and Duhon, '89) and phores and female cloacal glands that store Triturus (Halliday and Verrell, '84; Pecio, sperm (Sever, '94). Sever and Brunette ('93) '92) should be regarded as more plesiomor- noted that two main structural types of sper- phic than long-term sperm storage (several mathecae exist in salamanders. One type, weeks or more) documented (as reported called simple spermathecae, consists of nu- above) for some species. merous simple tubuloalveolar glands open- The total length of time over which Am- ing individually into the roof of the cloaca; phiuma tridactylum can store viable sperm simple spermathecae are found in all females was not determined in this study or in any in the suborder Salamandroidea except Pleth- previous study. The data indicate that the odontidae. Plethodontids are characterized potential for long-term sperm storage is pre- by complex spermathecae composed of com- sent, perhaps as long or longer than the 15 pound alveolar glands. No structural or devel- months suggested by Baker ('62). Some char- opmental reasons currently exist to consider acters mentioned below are no doubt related simple spermathecae homologous among the to the duration of sperm storage. families that possess them (Sever and Kloep- fer, '93). The identification and analysis of Staining reactions of the spermathecal characters associated with sperm storage in epithelium female salamanders may help resolve ques- Sever ('94) reviewed the literature on the tions concerning the evolution of sperma- reactions of spermathecal secretions to vari- thecae and the phylogeny of salamanders. ous stains at the light microscopy level and presented results for 71 species in which Duration of sperm storage in the spermathecae were stained with carbohy- spermathecae drate stains. As in the current study, the Cytologically and experimentally verified procedures used were the periodic acid- data are needed on the length of time sala- Schiff (PAS), which stains neutral carbohy- 94 D.M. SEVER ET AL. drates (such as glucose, mannose, galactose) tigrinum and N. beyeri are generally PAS+. and sialic acids, and alcian blue (AB) at pH In A. opacum, Sever and Kloepfer ('93) ob- 2.5, which stains primarily carboxylated gly- served both the electron-dense particle and cosaminoglycans (Kiernan, '90). In the Am- the flocculent material released into the lu- phiumidae, Sever ('94) reported that the sper- men during sperm maintenance, but in A. mathecae ofA. tridactylum, as verified herein, tigrinum Sever ('95)reported that the floccu- was PAS+, but that of A. pholeter was both lent material resulted from dissociation of PAS + and AB +. In other species with simple the electron-dense particle, and only the vacu- spermathecae, the staining reactions were ole containing the dissociated material was AB+ and PAS- in 25 of 28 species, including released (by exocytosis) into the lumen (dur- all 9 species of Ambystoma and 13 species of ing oviposition). Salamandridae examined (Sever, '94). Thus, In Amphiuma tridactylum, the fate of the A. tridactylum may be relatively unusual electron-dense particle of the secretory vacu- among salamanders with simple sperma- ole is unknown, but perhaps it dissociates thecae by having a PAS+ secretion. Closer into the finer product, as suggested by Sever examination of histochemical reactions, how- ('95)for the secretory product of Ambystoma ever, may reveal the presence of both neutral tigrinum. Release of the secretory product in and acidic carbohydrates in the sperma- Amphiuma tridactylum is apocrine, which so thecae of most if not all salamanders, al- far is unique among salamanders; the pro- though one type clearly may dominate. For cess has been described as merocrine in other example, Sever ('95) reported a PAS+ secre- species. Again, the phyletic significance of tion in Ambystoma tigrinum, and Sever et al. the secretory process cannot be determined. ('96) found PAS+ secretions in the salaman- Functions of the spermathecal secretions drid Notophthalmus viridescens. The inten- sity of staining reactions to the dominant In Ambystoma tigrinum, Sever ('95) found type of secretion may obscure reactions to that the spermathecal secretions were not other types of compounds, especially when released concomitant with the appearance of using multiple stains on the same tissue sperm in the spermathecae but with the act sample. of oviposition. He proposed, therefore, that the secretions serve to help flush sperm from Secretory vacuoles of varying or the spermathecae as eggs pass through the uniform density cloaca (contraction of myoepithelial cells has In the few species whose spermathecal se- also been implicated in this process [Hardy cretions have been studied by TEM, a di- and Dent, '871). Sever ('95)believed that the chotomy in the appearance of the secretory flushing of sperm is an ancestral state for vacuoles is apparent. In the plethodontid Eu- spermathecal secretions in which duration of rycea cirrigera (Sever, '91a) and in the sala- sperm storage was short. Providing the condi- mandrid Notophthalmus viridescens (Dent, tions for maintenance of sperm could be a '70; Sever et al., '96), the secretory vacuoles derived state to be searched for in species in are uniformly electron dense. In the ambysto- which sperm are retained for extended peri- matids Ambystoma opacum (Sever and Kloep- ods. Maintenance could involve either nour- fer, '93; Sever et al., '95) and A. tigrinum ishment (Benson, '68; Boisseau and Joly, '75) (Sever, '951, the proteid Necturus beyeri or providing the environment for sperm qui- (Sever and Bart, in press), and the salaman- escence (Hardy and Dent, '86). drid Salamandrina terdigitata (Brizzi et al., In Amphiuma tridactylum, the secretory '89, '951, the secretory vacuoles consist of an product was being produced during the mat- electron-dense particle surrounded (although ing season (March-May), but release of the not always centered) by a more electron- product into the lumen was not observed at lucent area called the "flocculent material" that time. Secretory material was found in by Sever and Kloepfer ('93). This is the type the lumen only in those specimens sacrificed of secretory vacuole we found in Amphiuma in August, and the product was often associ- tridactylum. Whether the solid or the vary- ated with degeneration of sperm. We did not ing secretory vacuole is more plesiomorphic observe the spermathecae of any females in a cannot be resolved with current data. postovipository state. The gravid females sac- The staining reaction of the vacuole seems rificed in August had presumably retained independent of the density of the vacuole, eggs through the normal ovipository period, since, for example, vacuoles of varying den- and, thus, sperm storage in these individuals sity in A. opacum are AB+ while those of A. may not have been following the pattern SPERM STORAGE IN AMPHIUMA 95 found in the natural environment. The non- in some aspects from sperm observed in the gravid specimens that contained sperm, how- testes, vasa deferentia, or spermathecae of ever, also showed signs of release of the secre- other salamanders (cf. Picheral, '79; Sever, tory material and degeneration of sperm in %la, '95). As opposed to other salamanders, the lumen. The role of the secretory product the axial fiber in the middle piece of the tail in in A. tridactylum still needs elucidation, but transverse section has two horseshoe-shaped a hypothesis of a function in sperm degenera- ends connected by a slender filament, and all tion is hereby tendered, which would be the portions are ensheathed by mitochondria. In first proposal of this function for spermathe- other species of salamanders, only one horse- cal secretions. shoe-shaped end is present in this region. More work on sperm ultrastructure in sala- Occurrence of spermiophagy within the manders is necessary to determine whether spermathecae such variation has functional or phyletic sig- Spermiophagy by the spermathecal epithe- nificance. lium was suggested in the newt Notophthal- In summary, we have studied some aspects mus viridescens by Dent ('70) but conclu- of sperm storage in female A. tridactylum, sively demonstrated first in salamanders in especially during the mating period, but much the plethodontid Eurycea cirrigera by Sever more work is needed before the process is ('91b, '92b) and Sever and Brunette ('93). fully understood in this enigmatic species. Sever and Kloepfer ('93) then reported on Efforts to induce lamper eels to mate under spermiophagy by the spermathecal epithe- laboratory conditions have been totally unsuc- lium of Ambystoma opacum, and Brizzi et al. cessful, and oviposition has been observed ('95) followed with a report of spermiophagy only once (Baker, '37; Kammeraad, '42). in the spermathecae of the salamandrid Sala- Thus, we still lack any data on the maximal mandrina terdigitata. Finally, Sever et al. duration of time between last mating and the ('96) extended the observations of Dent ('70) oviposition of a fertile clutch of eggs. Al- and confirmed that spermiophagy does in- though the species has proven to be a frustrat- deed occur in the spermathecae of N. uiride- ing one on which to conduct research on scens. Spermiophagy may be initiated early reproductive biology, additional data on in the breeding season before oviposition be- sperm storage, especially after fertilization of gins but ultimately results in the elimination a clutch, would be most worthwhile, as it of excess sperm within 2-3 months following would help address some of the issues raised oviposition (Sever, '92b; Sever and Kloepfer, in this paper concerning the functional and '93; Sever et al., '96). However, TEM studies phylogenetic significance of sperm storage in of two salamanders, Ambystoma tigrinum salamanders. (Sever, '95) and Necturus beyeri (Sever and ACKNOWLEDGMENTS Bart, in press), failed to find evidence of spermiophagy before or after oviposition, so This work received support from National the phenomenon may not be universal or Science Foundation grant DEB-9024918 to may simply need more careful study. D.M.S., although latter stages of the research The evidence for spermiophagy in Am- lacked support from NSF. This paper is pub- phiuma tridactylum is limited. Portions of lication 10 from the Saint Mary's College sperm were occasionally seen embedded in Electron Microscopy Facility. the spermathecal epithelium in all seasons, LITERATURE CITED but the only instance in which the sperm Adams, D.E. (1940) Sexual conditions in Triturus uirzde- certainly were deteriorating was in a non- wens 111. The reproductive cycle of' the adult aquatic gravid specimen sacrificed in August, in which form of both sexes. Am. J. hat.66:235-275. abnormal appearing sperm was found in epi- Armstrong, J.B., and S.T. Duhon (1989) Induced spawn- thelial cells along the luminal border and in ings, artificial insemination, and other genetic manipu- lations. In J.M. Armstrong and G.M. Malacinski (eds): cells that seemingly had broken free into the Developmental Biology of the Axolotl. London: Oxford lumen (Fig. 6C). Degradation of sperm in University Press, pp. 228-235. association with secretions into the lumen is Baker, C.L. (1945) The natural history and morphology more prevalent. ofAmphiuma. Rep. Reelfoot Lake Biol. Sta. 955-91. Baker, C.L. (1962) Spermatozoa of Amphiumae: Sper- mateleosis, helical motility and reversibility. J. Tenn. Sperm ultrastructure Acad. Sci. 37:23-37. Baker, C.L., L.C. Baker, and M.F. Caldwell(1947) Obser- The ultrastructure of the sperm stored in vation of copulation in Amphiuma tridactylum. J. Tenn. the spermathecae of A. tridactylum differed Acad. Sci. 22:87-88. 96 D.M. SEVER ET AL.

Baker, L.C. (1937) Mating habits and life history of Jordan, E.O. (1893) The habits and development of the Amphiuma tridactylum Cuvier and effect of pituitary newt (Diemyctylus uiridescens). J. Morphol. 8t269- injections. J. Tenn. Acad. Sci. 20:206-218. 366. Baylis, H.A. (1939) Delayed reproduction in the spotted Kammeraad, A. (1942) “Induced ovulation” in Am- salamander. Proc. Zool. SOC.,London 109At243-246. phiuma. Proc. Sac. Exp. Biol. Med. 49t195-196. Benson, D.G. (1968) Reproduction in urodeles 11. Obser- Kiernan, J.A. (1990) Histological and Histochemical vations on the spermatheca. Experientia 24t853. Methods: Theory and Practice, 2nd ed. Oxford: Per- Boisseau, C., and J. Joly (1975) Transport and survival of gamon Press. spermatozoa in female Amphibia. In E.S.E. Hafez and Kreeger, F.R. (1942) The cloaca of the female Amphiuma C.G. Thibault (eds): The Biology of Spermatozoa: tridactylum. Copeia 1942:240-245. Transport, Survival, and FertilizingAbility. Basel, Swit- Marynick, S.P. (1971) Long term storage of sperm in zerland: Karger, pp. 94-104. Desmognathus fuscus from Louisiana. Copeia 1971: Brimley, C.S. (1910)Records of some reptiles and amphib- 345-347. ians from Florida. Proc. Biol. SOC.Wash. 23:9-18. Massey, A. (1990) Notes on the reproductive ecology of Brimley, C.S. (1939) The amphibians and reptiles of red-spotted newts (Notophthalmus uiridescens). J. Her- North Carolina. Carolina Tips 2.67. petal. 24;106-107. Brizzi, R., G. Delfino, and C. Calloni (1989) Female McGregor, J.H. (1899) The spermatogenesis of am- cloaca1 anatomy in the spectacled salamander, Sala- phiuma. J. Morphol. 15(Suppl)t57-104. mandrina terdigitata (Amphibia: Salamandridae). Her- Parker, M.V. (1937) Some amphibians and reptiles from petologica 45310-322. Reelfoot Lake. J. Tenn. Acad. Sci. 1250-86. Brizzi, R., G.Delfino, M.G. Selmi, and D.M. Sever (1995) Pecio, A. (1992) Insemination and egg-laying dynamics in The spermathecae of Salamandrina terdigitata (Am- the smooth newt, Triturus uulgaris, in the laboratory. phibia: Salamandridae): Patterns of sperm storage and Herpetol. J. 2:5-7. degradation. J. Morphol. 223r2 1-33. Picheral, B. (1979) Structural, comparative, and func- Cagle, F.G. (1948) Observations on a population of the tional aspects of spermatozoa in urodeles. In D.W. salamander Amphiuma tridactylum Cuvier. Ecology Fawcett and J.M. Bedford (eds): The Spermatozoon 29t479490. Maturation, Motility, Surface Properties and Compara- Conant, R., and J.T. Collins (1991) A Field Guide to tive Aspects. Baltimore: Urban and Schwarzenberg. Reptiles and Amphibians Eastern and Central North Pool, T.B., and T.R. Hoage (1973) The ultrastructure of America. Boston: Houghton Mifflin Co. secretion in the spermatheca of the salamander, Man- Davison, A. (1895) A contribution to the anatomy and culus quadridigitatus (Holbrook). Tissue Cell 5t303- phylogeny ofAmphiuma means (Gardner).J. Morphol. 313. 11:375410. Rose, F.L. (1967) Seasonal changes in lipid levels of the Dent, J.N. (1970) The ultrastructure of the spermatheca salamander Amphiuma means. Copeia 1967r662-666. in the red spotted newt. J. Morphol. 132:397-424. Sever, D.M. (1991a) Comparative anatomy and phylog- Duellman, W.E., and L. Trueb (1986) Biology ofhmphib- eny of the cloacae of salamanders (Amphibia: Cau- ians. New York: McGraw-Hill. data). I. Evolution at the family level. Herpetologica Dundee, H.A., and D.A. Rossman (1989) The Amphibians 47t165-193. and Reptiles of Louisiana. Baton Rouge: Louisiana Sever, D.M. (1991b) Sperm storage and degradation in State University Press. the spermathecae of the salamander Eurycea cirrigera (Green).J. Morphol. 210:71-84. Fontenot, C.L., Jr. (1990) Sexual dimorphism and repro- Sever, D.M. (1992a) Comparative anatomy and phylog- ductive cycles in the aquatic salamander Amphiuma trzdactylum, in central Louisiana. M.S. thesis, South- eny of the cloacae of salamanders (Amphibia: Cau- eastern Louisiana University, Hammond. data). 111. Amphiumidae. J. Morphol. 211t63-72. Sever, D.M. (1992b) Spermiophagy by the spermathecal Halliday, T.R., and P.A. Verrell (1984) Sperm competi- epithelium of the salamander Eurycea cirrigera. J. tion in amphibians. In R.L. Smith (ed):Sperm Compe- Morphol. 212t281-290. tition and the Evolution of Mating Systems. New York: Sever, D.M. (1994) Observations on regionalization of Academic Press, pp. 487-508. secretory activity in the spermathecae of salamanders Hardy, M.P., and J.N. Dent (1986) Transport of sperm and comments on phylogeny of sperm storage in female within the cloaca of the female red-spotted newt. J. salamanders. Herpetologica 50r383-397. Morphol. 190r259-270. Sever, D.M. (1995) Spermathecae of Ambystoma tigri- Hardy, M.P., and J.N. Dent (1987) Hormonal facilitation num (Amphibia Caudata): Development and a role for in the release of sperm from the spermatheca of the the secretion. J. Herpetol. 29r243-255. red-spotted newt. Experientia 43:302-304. Sever, D.M., and H.L. Bart, Jr. (in press) Ultrastructure Hay, O.P. (1888) Observations on Amphiuma and its of the spermathecae of Necturus beyeri (Amphibia: young. Am. Nat. 22t315-321. Proteidae) in relation to its breeding season. Copeia. Hildebrand, C.E. (1910) A nest of the congo eel, Am- Sever, D.M., and N.S. Brunette (1993) Regionalization of phiuma means. Proc. Biol. Sac. Wash. 23:lO. eccrine and spermiophagic activity in the spermathe- Houck, L.D., and K. Schwenk (1984) The potential for cae of the salamander Eurycea cirrigera (Amphibia: long-term sperm competition in a plethodontid sala- Plethodontidae). J. Morphol. 21 7:161-170. mander. Herpetologica 40r410415. Sever, D.M., and N.M. Kloepfer (1993) Spermathecal Houck, L.D., and S.K. Woodley (1995) Field studies of cytology of Ambystoma opacum (Amphibia: Ambysto- steroid hormones and male reproductive behavior in matidae) and the phylogeny of sperm storage organs in amphibians. In H. Heatwole (ed): Amphibian Biology, female salamanders. J. Morphol. 21 7t115-127. Vol. 2: Social Behaviour. Chipping Norton, Australia: Sever, D.M., J.D. Krenz, K.M. Johnson, and L.C. Rania Surrey Beatty and Sons, pp. 677-703. (1995) Morphology and evolutionary implications of Humphrey, R.R. (1977) Factors influencing ovulation in the annual cycle of secretion and sperm storage in the Mexican axolotl as revealed by induced spawnings. spermathecae of the salamander Ambystoma opacum J. Exp. Zool. 199:209-214. (Amphibia: Ambystomatidae). J. Morphol. 223:35-46. SPERM STORAGE IN AMPHIUMA 97

Sever, D.M., L.C. Rania, and J.D. Krenz (1996) The Wilbur, H.M. (1977)Propagule size, number, and disper- annual cycle of sperm storage in the spermathecae of sion pattern in Ambystoma and Asclepias. Am. Nat. the red-spotted newt, Notophthalmus viridescens (Am- 111:43-68. phibia: Caudata). J. Morphol. 227:155-170. Wilson, F.H. (1940) The life cycle of Amphiuma in the Sturdivant, H.P. (1949)The sperm cycle in Amphiuma. vicinity of New Orleans based on a study of the gonads J. Tenn. Acad. Sci. 24:170. and gonoducts. Anat. Rec. 78(Suppl.): 10P105 (ab- Trauth, S.E. (1983) Reproductive biology and spermathecal anatomy of the dwarf salamander (Eurycea stract). quadrzdigitata) in Alabama. Herpetologica 39:9-15. Wilson, F.H. (1941) The cloaca in the male Amphiuma Ultsch, G.R., and S.J. Arceneaux (1988)Gill loss in larval tridactylum. Anat. Rec. 81(Suppl.):63 (abstract). Amphiuma trzdactylum. J. Herpetol. 22:347-348. Wilson, F.H. (1942)The cycle of egg and sperm produc- Weber, J.A. (1944) Observations on the life history of tion in Amphiuma tridactyla Cuvier. Anat. Rec. Amphiuma means. Copeia 1944:61-62. 84(Suppl.):532(abstract).