Introduction
The process of speciation is frequently accompanied by divergence in life history traits, due to different environmental pressures acting on the newly-forming species (Abbott et al. 2013, Doellman et al. 2018). Divergence in these phenotypic traits can be induced by environmental conditions or can be genetically determined (Fischer et al. 2016, Le Sage et al. 2021, Sabino-Pinto et al. 2019). Genetically-determined traits that increase population fitness of a species in its environment are called adaptive traits, and the process by which trait phenotypes and underlying genotypes increase in frequency in a population is referred to as local adaptation (Hereford 2009, Hoban et al. 2016, Savolainen et al. 2013). Adaptive traits may enforce the process of speciation as hybrids often exhibit intermediate values of such traits, which are associated with lower fitness of the hybrids in either parental habitat (Barton 2001, Delmore and Irwin 2014, Hatfield and Schluter 1999, McGee et al. 2015, Seehausen 2004). On the other hand, phenotypic plasticity, allows the expression of different phenotypes from a single genotype due to differing environmental stimuli (Benard 2004, Fitzpatrick 2012, Michimae 2006, Pfennig et al. 2010, Van Buskirk 2009). Plastic traits generally do not decrease fitness of hybrids, which can lead to genetic homogenization of the newly-forming species and hinder lineage divergence (Abbott et al. 2013, Gow et al. 2006, Nonaka et al. 2015, Seehausen 2006, Taylor et al. 2006). As both plastic and genetically determined traits have an effect on speciation and play a role in the ability of populations to thrive in their environments, it is crucial to understand if traits delimiting closely related species are plastic or genetic in nature (Campbell‐Staton et al. 2018, Corl et al. 2018). Such knowledge is central to our ability to 1) delimit species and conservation management units, 2) inform management strategies with respect to conservation priorities and reintroductions, and 3) better understand the magnitude of environmental change that any population can tolerate.
Ambystoma barbouri and A. texanum are considered two separate, closely related salamander species (Kraus and Petranka 1989). The two salamanders are generally differentiated by traits associated with their breeding behavior. Ambystoma barbouri is typically a stream breeding salamander, which attaches its eggs in a single layer underneath flat (typically limestone) rocks of first and second order streams (Drayer et al. 2020, Holomuzki 1991, Kats and Sih 1992, Niemiller et al. 2009). Ambystoma texanum is typically a vernal pool breeder that deposits eggs in small clumps on submerged vegetation, as is typical for the genus Ambystoma (Petranka and Sih 1987, Pfingsten et al. 2013). The differences in reproductive traits between them have been used to delimitate their ranges, as the two species are morphologically similar (Kraus and Petranka 1989 but see Petranka 1989 on differences in tooth morphology). Based on this delimitation, the two species are generally considered parapatric. Ambystoma barbouris range extends from the Nashville basin to central Ohio (Figure 1). Due to its small range, A. barbouri has been listed as near-threatened in Tennessee and has recently been placed under a Species Status Assessment for its full range. In contrast to the limited range of A. barbouri , A. texanum inhabits an expansive range from central Texas to northern Ohio and southern Michigan (Figure 1).
The reproductive and larval traits such as oviposition location, egg size, clutch size, and larval pigmentation that differentiate A. barbouri from A. texanum (Petranka 1982, Petranka and Sih 1987, Venesky and Parris 2009) seem to exhibit a parapatric distribution. In some locations an abrupt switch in reproductive traits can be observed within kilometers, with no intermediate forms observed (Kraus and Petranka 1989). For example, the area running from western Kentucky to central Ohio represents a presumed zone of sympatry, as breeding habitats are not clearly differentiated and both species can be found in non-traditional habitats such as roadside ditches (Pfingsten et al. 2013). Intermediate forms of reproduction-related traits from this sympatric zone, such as clutch size and larval size, have not been documented in the literature (Petranka 1982).
The genetic structure of the two species is also complex. In some parts of their distribution, the geographic boundary between stream- versus pool-related reproductive traits seem to correspond well with the geographic distribution of two divergent mitochondrial DNA clades (Bi and Bogart 2010, Denton et al. 2014, Eastman et al. 2009, Williams et al. 2013). However, this pattern is not exclusive, as signals of mitochondrial introgression and mito-nuclear discordance have been documented, with some A. barbouri mtDNA genomes present withinA. texanum range and breeding habitat (Denton et al. 2014). These results also suggest that the two taxa might not be reproductively isolated but that selection pressures associated with differing breeding habitats might have resulted in species-specific or population-specific adaptations. This raises a question to what extent the reproduction-related traits are adaptive or plastic, whether barriers to hybridization exist between the two species, and whether potential hybrids exhibit inferior fitness.
The objective of this research was two-fold. First, we compare reproduction-related traits of A. barbouri  and A. texanum  females, mated with either conspecific or heterospecific males in laboratory settings. We hypothesized that the traits associated with their distinct reproductive strategies/habitats, such as egg deposition location or clutch size, are genetically-determined rather than phenotypically plastic and represent adaptations to the species’ different breeding habitats. If true, we expected that the females of the two species would maintain their distinct reproductive traits when bred under a common treatment (Richter-Boix et al., 2013). Second, we compared life-history traits of hybrid versus nonhybrid offspring (Hall, 1999). We hypothesized that hybridization has negative consequences which would potentially explain the geographically abrupt turnover of reproductive and developmental traits observed in the A. barbouri / A. texanum  sympatric zone. If hybrid inferiority was present, we expected to find one of the following: 1) breeding between heterospecific pairs to be less frequent than breeding between conspecific pairs, 2) the viability of hybrid eggs or larvae to be reduced, or 3) hybrid offspring to exhibit intermediate phenotypes in comparison to non-hybrid offspring making them less suitable for either parental habitat.