(Murphy, 1962) to hundreds (Graham, with males exhibiting nudging, head- B. Eggs. 1971; Shoop and Doty, 1972; Stenhouse, swinging, lifting, and body-flexing be- i. Egg deposition sites. Breeding sites are 1985a), ϳ1,000 (Pechmann et al., 1991; haviors (Arnold, 1972). Spermatophore generally the dried beds of temporary Semlitsch et al., 1996) to Ͼ10,000 (Taylor deposition follows lateral undulations of ponds, the margins of reduced ponds, or and Scott, 1997). However, given the re- the tail. Spermatophores are 4–5.5 mm tall dry floodplain pools. Female marbled liance of marbled salamanders on small (Lantz, 1930; illustrated in Noble and salamanders construct nests and lay eggs isolated seasonal wetlands and intact Brady, 1933). Typical and secondary sper- under virtually any cover in situations forested floodplain habitats, their abun- matophore deposition may occur (Arnold, where the nest is likely to be flooded by dance presumably has declined as wetland 1972, 1976; personal observation); a male subsequent winter rains (Noble and Brady, habitats have been destroyed (Petranka, may deposit over 10 spermatophores in 1933). Eggs are laid on the edges of pools 1998). For example, from the 1950s–70s 30–45 min (L. Houck, personal communi- (Dunn, 1917b) and in dry basins under the loss of wetlands in the Southeast was cation). Males will mate with females vegetation ( Jackson et al., 1989), logs greater than in any other region of the outside what is typically thought of as (Bishop, 1924; Doody, 1996), and leaf de- country, with a net annual loss of 386,000 the wetland margin (Krenz and Scott, bris (Deckert, 1916; Petranka and Petranka, ac/yr (Hefner and Brown, 1985); in North 1994). Males often will court other males 1981b). Eggs are laid occasionally on non- Carolina approximately 51% of all wet- (Noble and Brady, 1933), including sper- soil substrate (Brimley, 1920a). Nest site land acreage on the Coastal Plain has matophore deposition in the absence of a selection by females is influenced by mi- been lost (Richardson, 1991), including female (L. Houck, personal communica- crosite elevation within the pond bed, 70% of the pocosins that have been “de- tion). Females may follow a male to pick site hydrologic regime, cover availability, veloped” or partially altered (Richardson, up a spermatophore (Noble and Brady, and soil moisture (Petranka and Petranka, 1983); approximately 97% of the Carolina 1933) or simply move about an area until 1981a,b; Jackson et al., 1989; Figiel and bays on the Coastal Plain of South Carolina a spermatophore is located (Arnold, Semlitsch, 1995; Wojnowski, 2000; but have been altered or severely impacted; 1972). When a spermatophore contacts a see also Marangio and Anderson, 1977). and Ͻ200 bays of the original thousands female’s vent she will lower herself onto it Females actively excavate oblong to ovoid- are “relatively unimpacted” (Bennett and and insert it into her cloaca (Arnold, shaped depressions (King, 1935; Petranka Nelson, 1991). 1972). Sperm are stored in exocrine glands and Petranka, 1981b). called spermathecae in the roof of the ii. Clutch size. Of the three reproductive 3. Life History Features. cloaca (Sever and Kloepfer, 1993). Eggs are modes of salamanders outlined by Salthe A. Breeding. Reproduction is terrestrial, fertilized internally by sperm released (1969), marbled salamanders have an in or near the wetland breeding sites prior from spermathecae during oviposition atypical type I mode (Salthe and Mecham, to pond filling. (Sever et al., 1995). Females may pick up 1974; Kaplan and Salthe, 1979). Clutch size i. Breeding migrations. Onset of breeding multiple spermatophores (Arnold, 1972), ranges from approximately 30 to Ͻ200 migrations occurs from September–No- but sperm competition has not been de- eggs (see Petranka, 1998) and usually is vember. Timing varies geographically and finitively demonstrated. Sperm in the positively correlated with female body may occur 1 mo or more earlier at south- spermathecae do not persist for 6 mo size (Kaplan and Salthe, 1979; Walls and ern latitudes compared with northern lat- after oviposition (Sever et al., 1995). Males Altig, 1986; Petranka, 1990; Scott and Fore, itudes (Anderson and Williamson, 1973). will mate with females beyond what hu- 1995), although not always (Kaplan and On a broad scale, seasonal migrations are mans typically define as the wetland mar- Salthe, 1979)—larger females may have probably linked to regional climatic and gin (Krenz and Scott, 1994). larger eggs (Kaplan and Salthe, 1979). hydrological cycles (Salthe and Mecham, ii. Breeding habitat. Marbled salaman- Compared to other species of Am- 1974). Adult salamanders move to breed- ders are one of two species of Ambystoma bystoma, females in some populations of ing sites on rainy nights and tend to enter that breed on land (Petranka, 1998), and marbled salamanders may have fewer, and exit the site at approximately the they are the only Ambystoma species that larger eggs than would be expected for an same point (Shoop and Doty, 1972; P.K. exhibit parental care (Nussbaum, 1985, animal of their size (Kaplan and Salthe, Williams, 1973; Douglas and Monroe, 1987). Due to the terrestrial reproductive 1979; D.E.S., unpublished data; for a dif- 1981; Stenhouse, 1985a). Males generally habits of marbled salamanders, breeding ferent opinion see Nussbaum, 1985, 1987). arrive at the breeding site before females is restricted to fish-free wetlands with sea- For example, comparing female marbled (Noble and Brady, 1933; Graham, 1971; sonally fluctuating water levels that in- salamanders and mole salamanders (A. Krenz and Scott, 1994). In a 25-yr study at clude upland hardwood “swamp forests” talpoideum) of equal body size, marbled Rainbow Bay in South Carolina (see Seml- (King, 1935), bottomland hardwood pools salamanders have 3–4 times fewer eggs, itsch et al., 1996), the mean date of arrival (Viosca, 1924a; Petranka and Petranka, but each egg is 3–4 times larger with 3–4 of males at the breeding site was 10 d ear- 1981a,b), quarries (Graham, 1971), vernal times higher lipid amounts (Komoroski, lier than females (unpublished data), per- ponds (Doty, 1978), Carolina bays ( Jack- 1996; D.E.S., unpublished data). Mean haps due solely to the combination of a son et al., 1989; Gibbons and Semlitsch, egg diameter is greater in marbled sala- limited number of nights suitable for mi- 1991), and floodplain pools (Petranka, manders than in flatwoods salaman- gration and slower nightly movements by 1990). Females remain with eggs (Noble ders (A. cingulatum; 2.8 vs. 2.3. mm; fat, gravid females (Blanchard, 1930; per- and Brady, 1933) for varied lengths of Anderson and Williamson, 1976). Mean egg sonal observations). The sex ratio of the time (Petranka, 1998); they may leave be- dry mass is greater in marbled salaman- breeding population is often biased to- fore eggs are inundated (McAtee, 1933; ders than in either mole salamanders ward males (Graham, 1971; Stenhouse, Jackson et al., 1989; Petranka, 1990). Nest or spotted salamanders (A. maculatum; 1987; Krenz and Scott, 1994), in part be- brooding appears to enhance embryonic Komoroski, 1996). The caloric content cause males mature at an earlier age survival (Petranka and Petranka, 1981b; (cal/mg dry mass) of marbled salamander (Scott, 1994; Pechmann, 1994). The sex Jackson et al., 1989), although the mecha- eggs is greater than the energy content of ratio in one study (Parmelee, 1993) during nism is unknown. Opinions differ on spotted salamanders and tiger salamander the non-breeding season did not differ whether there is an energetic cost to (A. tigrinum) eggs (Kaplan, 1980b). Rela- from 1:1. brooding by females (Kaplan and Crump, tively few, large eggs with lipid stores in a. Courtship activity. At the time of au- 1978; versus D.E.S., unpublished data). excess of the amount needed for embryo- tumn migration, males are at maximal Occasionally nests are communal (Gra- genesis probably reflects a response to the testosterone levels (Houck et al., 1996; un- ham, 1971; Petranka, 1990), especially if terrestrial breeding habits of marbled sala- published data). Courtship is terrestrial, cover items are scarce (Palis, 1996b). manders and the extreme variability and 628 AMBYSTOMATIDAE unpredictability in the timing of hatching 1996; table 1), depending upon their den- 1970). Larvae nearing metamorphosis re- (i.e., the duration of the egg stage). Sub- sity, food levels, and temperature. Growth main near the bottom at night (Petranka stantial (15–30%) variation in egg diame- rates are comparable to spotted salaman- and Petranka, 1980). ter occurs within and among populations ders (Walls and Altig, 1986), but compar- iii. Larval polymorphisms. None reported, (Kaplan, 1980a). Egg size is positively isons to mole salamanders differ (Keen et although behavior differences are known. correlated with hatchling size and early al., 1984; Walls and Altig, 1986). Although Laboratory assays have demonstrated two larval size (Kaplan, 1980a). larval growth is temperature dependent divergent aspects of kin recognition. In In spite of terrestrial egg laying, egg (Stewart, 1956), temperature effects may some contexts, kin recognition may reduce structure in marbled salamanders is simi- not be as pronounced as in some other aggression and cannibalism among sib- lar to aquatic Ambystoma species (Salthe, Ambystoma species (Keen et al., 1984). lings in larval marbled salamanders (Walls 1963). Egg development is temperature- Food level, temperature, hatching time, and Roudebush, 1991); whereas in other dependent (Noble and Brady, 1933); de- and larval density affect traits of newly contexts, large larvae may eat siblings velopment (at similar temperatures) is metamorphosed animals (Stewart, 1956; preferentially (Walls and Blaustein, 1995). slower than for some other ambystom- Boone et al., 2002). Early hatching larvae Hokit et al. (1996) further demonstrated atids (Moore, 1939). The prospective neu- are larger at metamorphosis, have higher that kin discrimination is context depend- ral tissue of marbled salamanders has a survival, and metamorphose earlier than ent.