Derived monomorphism is a condition in which males and females are phenotypically similar, but the similarity is derived.  Derived monomorphism typically evolves from sexual dimorphism or from a different monomorphic state.  We examined the hormonal basis of derived monomorphism in the salamander genus Aneides (Plethodontidae).  We reject our hypothesis that circulating levels of androgens explain the derived traits, such as enlarged jaw musculature, in females (some would call them “male-like traits”).  There was no clear pattern of differences in androgen levels or degree of dimorphism in androgen levels, between the sexually dimorphic Aneides hardii and the other, derived monomorphic, species studied.  Concentrations of testosterone and dihydrotestosterone were higher in males than in females in all species examined. The degree of sexual dimorphism in androgen level was also consistent among the species studied. Levels of androgens in female plethodontids have not been previously reported.

Sample collection

Individuals of A. lugubris, A. ferreus, A. flavipunctatus, A. hardii, and P. elongatus were hand collected in the field (see Table 1 for sample sizes and Appendix for localities and dates), sacrificed by decapitation within 3 minutes of capture, and blood samples collected in heparinized capillary tubes and stored on ice.  Within three hours of collection, blood samples were centrifuged and the plasma frozen.  Samples were kept at –80°C until radioimmunoassays (RIAs) were performed.  Animals were fixed in 10% neutral buffered formalin, rinsed in water, and stored in 70% ethanol.  Sex was determined via gonad inspection.  One testis was removed from males for histological analysis.  Authorization for this work included: Oregon Dept of Fish and Wildlife permits 437-93, 051-94, 038-95, 045-96; New Mexico Dept of Game and Fish permit 1702 (1993-96); California Department of Fish and Game permits 4023 1993-95, 4755 1995-1997.  

Blood samples were collected across the animal’s active season: A. lugubris (October-May), A. ferreus (September – May), and A. hardii (May-August).  A. flavipunctatus samples were collected in March only.  For P. elongatus, samples were collected over several months (October, November, January, March), but sample sizes were too low to analyze seasonally.

Radioimmunoassays

Plasma Estradiol (E), testosterone (T), and dihydro-testosterone (DHT), were separated using column chromatography and measured by radioimmunoassay (RIA).  Extraction, Celite microcolumn chromatography, and radioimmunoassay techniques followed the protocol in Wingfield et al. (1982).  Basically, plasma samples were equilibrated overnight with 800-1000 cpm of each steroid hormone to be measured to determine recovery efficiency.   Plasma was extracted with diethyl ether, dried under nitrogen gas, and resuspended in 10% ethyl acetate in iso-octane.  Samples were chromatographed on microcolumns of Celite:propylene/ethylene glycol.  The eluate solvents used were 10% ethyl acetate in iso-octane for DHT, 20% ethyl acetate in iso-octane for T, and 40% ethyl acetate in iso-octane for E.  Column aliquots were dried under nitrogen and resuspended in phosphate buffer.  Two aliquots were taken (1, 1:20 for DHT, and T, and duplicates for E).  Aliquots were incubated overnight at 4C with the tritiated steroid to be measured and its respective antibody (E antibody, 17-94 from Endocrine Sciences, Tarzana, CA; androgen antibody, T3003). Following the incubation period, the assay tubes received a 5% charcoal: 0.5% dextran mixture to remove unbound steroid.  Tubes were vortexed, incubated for 15 min at 4°C, and centrifuged for 10 min at 3,000 rpm.  The supernatant was collected from each sample, equilibrated overnight, and counted in the Beckman scintillation counter the following day (see Navara et al. 2006, Wingfield et al. 1982).

Sensitivities were 15 pg/ml for T, DHT, and E. Inter-assay coefficients of variation were E: 13% (n=11), T: 10% (n=11), DHT: 8% (n=10).  Values for intra-assay variation were E: 8% (n=138), T: 7% (n=150), DHT: 8% (n=131).   

Histology

Using standard histological procedures, testes were embedded in paraffin, sectioned at 12 um, mounted on slides, and stained with hematoxylin and eosin (Presnell and Schreibman 1997). 

Statistical Analysis

To examine our hypothesis, we modeled sex steroid concentration as a function of sex and species for A. hardii, A. ferreus, A. flavipunctatus, A. lugubris, and P. elongatus (i.e. five-species model). Because little has been published on sex steroid variation in these species, we first modeled E, T, and DHT concentration as a function of month and sex, to determine which months to include in the five-species model. We first fit interaction models of month by sex for adult A. hardii, A. ferreus and A. lugubris. If interactive effects were not significant, we dropped the interaction term, refit and interpreted main effects models of month and sex. Our P. elongatus and A. flavipunctatus sample sizes were too small to fit sex by month models.

Following our interpretation of the sex by month models (see Results, Seasonal Variation), our five-species model included June, July, and August observations for A. hardii, January, February, March, and November observations for A. lugubris, and January, February, March, May, August, October, November, and December observations for A. ferreus, March observations for A. flavipunctatus, and January, March, and November observations for P. elongatus. Our hypothesis predicted we would observe higher levels of T and DHT in males than in females of the sexually dimorphic A. hardii, but higher levels of T and DHT in female A. ferreus, A. flavipunctatus, and A. lugubris, compared to female A. hardii, due to derived monomorphism. We therefore expected to observe a large sex by species interaction in the five-species model, with pairwise comparisons of sex within species revealing different levels of T and DHT in A. hardii, and less of a difference in A. ferreus, A. flavipunctatus, and A. lugubris.

Because little is known about sex steroid variation in these species, we also report results of modeling E, T, and DHT as a function of sex and species in immature A. hardii and A. ferreus. Sample sizes for A. lugubris immatures were too low to be included in the modeling analyses.

For all models, we included data with and without potential outliers, as suggested by Pollet and van der Meij (2017). We identified potential outlier observations as those more than 1.5 interquartile range units away from the median, within sex by month groups. Modeling results were equivalent with and without outliers, so we retained all observations for modeling. We log-transformed our hormone observations to stabilize variance and analyzed residuals to evaluate model fit.  We made post-hoc comparisons using the Tukey HSD test. We also provide summary statistics for individual T/DHT ratios by sex and species, and mean male/female T and DHT ratios by species. All analysis was done using R (R Core Team 2023).