Abstract
Hybrid zones are excellent areas to investigate evolutionary processes such as speciation, adaptation, and multiple stable states, and can also give insight into historical biogeography. Using 369 samples from 17 ponds, I examined mitochondrial DNA patterns across a hybrid zone between the tiger salamander subspecies Ambystoma tigrinum mavortium and A. t. tigrinum. The mtDNA cline was non-coincident relative to a series of overlapping clines of other characters, and is the third non-coincident cline found near this subspecies boundary. Evidence from biogeography, paleobotany, and coalescence theory, combined with cumulative data on the zone’s clinal patterns, suggest the subspecies evolved in allopatry, reestablished an interface in the southern portion of the current hybrid zone, then recently spread north together as the glaciers receded.

Introduction
The tiger salamander Ambystoma tigrinum is a common, polytypic species that ranges through most of the continental United States, stretching into Canada and the Mexican Plateau. Six subspecies within the complex are diagnosed primarily by adult color pattern, larval gill raker counts, and incidence of neotony (Collins et al., 1980). These characters, as will be discussed, may not reliably identify genetic differences (Dunn, 1940; Pierce & Mitton, 1980). Tiger salamanders inhabit forests, meadows, deserts and canyons throughout the year, and are easy to catch when they move to fishless ponds during the breeding season, or as aquatic larvae.

The species’ recent expansions following the last glacial maxima have moved their current distributions into previously glaciated areas where they have likely only recently stabilized (Pielou, 1991). Current subspecies distributions have no overlap, and instead border on each other with resulting hybrid zones that feature a wide range of subspecies-level interbreeding (Bogart & Klemmens, 1997; Pierce & Mitton, 1980). Hybrid zones are natural evolutionary microcosms. They offer insight into the crossroads of speciation and genetic fusion, and into the possibilities of multiple stable states. Studying differing patterns of introgression among loci helps us understand how evolution sorts genetic variation. Something these clinal patterns don’t tell us is how a zone came to exist. Did it emerge in situ, or through secondary contact (Kocher, 1986)?

Hybrid Zones
Many species exist in a patchwork of genetically distinct localized populations, which hybridize in transition zones. They have been argued to exist between populations at different multiple stable states of Wright’s shifting balance theory, or along environmental gradients. These are often termed “tension zones”, and result from either the balance between random dispersal and selection against hybrids (Barton et al., 1989; Woodruff, 1979) or positive assortative mating (Harrison, 1989).

Theory and data show that hybrid zones are attracted to areas of low population, which often coincide with physical barriers such as rivers (Barton et al., 1989; Hewitt, 1989). Theory also suggests hybrid zones move to minimize their length by smoothing out bulges, resulting in a somewhat equal width throughout (Hewitt, 1989). Hybrid zones themselves tend not to move unless dramatic environmental changes occur (Hewitt, 1989).

Transitions across hybrid zones are usually narrow relative to range and dispersal distance. Even apparently neutral alleles rarely penetrate far, although neutral alleles may be linked to genes under selective pressure (Hewitt, 1989). Zones can include any number of characters, but predominately involve numerous morphological, behavioral and genetic characters, the patterns of which tend to mimic each other (Hewitt, 1989).

Discordances: Morphological vs. Genetic Patterns
Although clines predominately feature overlapping characters, exceptions have been documented. Studies have shown discordances between morphological characters used to identify subspecies and genetic markers of relationships between them. Zink and Dittman (1993), for example, found no relationship among mtDNA haplotypes and the geographically localized phenotypic characters used to distinguish among subspecies of song sparrows. They suggested the morphological characters (plumage and size) evolved quicker that mtDNA monophyly because they are polygenic. Larson (1983) has also found discrepancies between diagnostic characters of the Bolitoglossa rufescens salamander group and robust genetic indicators of their evolutionary relationships.

Discordances: MtDNA vs. Nuclear Patterns
MtDNA and nuclear genes can have different patterns of introgression, phenomena often explained as resulting from founder effects in unpopulated areas (Gyllenstein & Wilson, 1987). In one case where populations throughout a hybrid zone have maintained large numbers for a considerable time, mtDNA mimics nuclear patterns (Wilson et al., 1985). These empirical results counter theory and computer simulations that predict that mtDNA can more easily spread across gene pools because of its lower effective population size and relative selective autonomy as an unlinked gene (Hewitt, 1989; Wilson et al., 1985; Barton & Hewitt, 1989; Harrison, 1989). Different patterns may, however, simply reflect the time genes have had to introgress (Vanlerberghe et al.; Marchant et al (as cited in Barton & Hewitt, 1989)).

Ambystoma tigrinum hybrid zones
Several hybrid zones between Ambystoma subspecies have been documented. Pierce and Mitton (1980) studied an interface between Ambystoma tigrinum nebulosum and A. t. mavortium located along the continental divide. Intermediate color and allozyme patterns overlapped each other in a cline through the zone. Salamanders within the zone suffered no apparent reduced fitness.

Kocher (1986) studied allozyme cline patterns in two Ambystoma hybrid zones. One was between A. t. tigrinum (the Eastern tiger salamander ) and A. t. mavortium (the Barred tiger salamander) along the southeastern border of Nebraska, and the other between A. t. mavortium and A. t. nebulosum in southern Colorado. The A. t. tigrinum - A. t. mavortium clines were non-coincident, and varied in width among loci; cumulatively they were nearly 125 kilometers wide. For most loci, clines were approximately 30 km, although clines for Mdh-2 and Pgi were abnormally wide. Most clines centered around 25 km west of the Missouri River Valley, in agreement with range maps. Two, however, deviated substantially: one (Aco-c) 40 km west, and the other (Mhd-2a) 35 km east. Kocher’s detailed survey showed hybrids were not selected against, and may in fact have an advantage.

Extreme morphological variation of tiger salamanders within and among populations throughout its range leads to problems with overlapping characters, as is the case with the Barred and Eastern subspecies, which we examined.

In this study I will characterize the hybrid zone between the barred and eastern subspecies based on mtDNA markers. I will compare these results to those based on morphological characters as well as other genetic markers. Comparisons and conclusions will be made in the light of a wealth of old and new data that strongly suggest the zone is both recent, and the outcome of secondary contact.