Background : The hybridizing field crickets, Gryllus firmus and Gryllus pennsylvanicus have several barriers that prevent gene flow between species. The behavioral pre-zygotic mating barrier, where males court conspecifics more intensely than heterospecifics, is important because by acting earlier in the life cycle it has the potential to prevent a larger fraction of hybridization. The mechanism behind such male mate preference is unknown. Here we investigate if the female cuticular hydrocarbon (CHC) profile could be the signal behind male courtship. Results : While males of the two species display nearly identical CHC profiles, females have different, albeit overlapping profiles and some females (between 15-45%) of both species display a male-like profile distinct from profiles of typical females. We classified CHC females profile into three categories: G. firmus -like (F; including mainly G. firmus females), G. pennsylvanicus -like (P; including mainly G. pennsylvanicus females), and male-like (ML; including females of both species). Gryllus firmus males courted ML and F females more often and faster than they courted P females (p<0.05). Gryllus pennsylvanicus males were slower to court than G. firmus males, but courted ML females more often (p<0.05) than their own conspecific P females (no difference between P and F). Both males courted heterospecific ML females more often than other heterospecific females (p<0.05, significant only for G. firmus males). Conclusions : Our results suggest that male mate preference is at least partially informed by female CHC profile and that ML females elicit high courtship behavior in both species. Since ML females exist in both species and are preferred over other heterospecific females, it is likely that this female type is responsible for most hybrid offspring production.

Collection

In August 2013, we collected penultimate instar G. firmus crickets from Guilford, CT (41˚.13’,-72˚40’) and G. pennsylvanicus crickets from Ithaca, NY (42˚25’,-76˚.29’), allopatric pure species populations. Individuals were separated by sex and species, and raised at room temperature (25˚C) in plastic cages (33×20×13 cm, with a maximum of 12 individuals) with ad libitum food (a mixture of Purina Cat Chow®, LM Bonanza Rabbit Food®, and Fluker’s Cricket Feed®) and water.

Courtship Trials

To measure male courtship intensity as a function of speed, we placed males in a petri dish with either a heterospecific or conspecific female and recorded the time to the start of courtship. We conducted four sets of crosses each day (10 AM and 2 PM). In each cross set (am and pm), a male was paired to a conspecific female followed by a heterospecific female an hour later (or reverse, heterospecific then conspecific). In total each male was placed with four conspecific and four heterospecific females over a two day period (alternating the order of conspecific and heterospecific females).  Females were also tested to eight males, but were kept virgin throughout the experiment. Based on previous work, (Maroja et al., 2014) we limited time to a maximum of 25 minutes (95% of males either already initiated or will not initiate courtship past this time), if the male did not initiate courtship within that time, the trial was considered unsuccessful. Pairs were never allowed to mate; after the male initiated courtship the pair was immediately separated. All male and female crickets used in the experiment were of approximately 12 days old (within 1-4 days apart); the adult lifespan in captivity is 30±8 days (personal observation).  We measured pronotum size as a proxy for body size in both males and females. and then compared species with a two-way ANOVA analysis.

Cuticular hydrocarbon analysis

We extracted CHC from all individuals used in the courtship experiment by placing whole crickets into glass vials containing 3mL (females) or 2mL (males) of HPLC grade hexane for 5-7 minutes. For the analysis we transferred the CHC samples into 2mL clear glass surestop vials with 300µL glass inserts and analyzed with Agilent Technologies (AT) 7890A GC system with an (AT) HP-5ms (325˚C 30m x 250 µm x .250 µm) column attached to an AT 5975C inert XL EI/CI MSD with triple-Axis Detector MS system, that obtains chromatograms and both electron and chemical ionization mass spectra. The GCMS method consisted of a 2 µL of each sample injected in a split mode with a split ratio of 100:1. The column was held at an initial temperature of 100˚C for 1 minute followed by 15˚C/min increase to 180˚C, then a 3˚C/min increase to 260˚C, and finally a 1˚C/min increase to a final temperature of 280˚C held for 10 minutes.

For GCMS data analysis, we scored a total of 17 peaks for each individual.  Ten of these peaks were previously used as representative of common compounds in males and females (Maroja et al., 2014) and we also scored seven new peaks. To score the peaks as a relative proportion of the total, we took the percent area contributed by each of the scored peaks and scaled the scored peaks to add up to 100% for each individual. To account for the dependence in the relative proportions, we used the centered log ratio (CLR) transformation prior to further analyzing the relative proportions of the peaks (Aitchison, 1982).

Cuticular hydrocarbon data was visualized via principal component analysis (appendix figure 1). Then the female CHC data were clustered into homogenous groups, or CHC categories, based on the Euclidean distance of CLR transformed relative proportions of the 17 peaks using the partitioning around medoids algorithm (Kaufman & Rousseeuw, 1990). We chose the number of groups that maximized the average silhouette, a cluster validity measure which measures the cohesion and separation of the clusters (Rousseeuw, 1987). This clustering process of female CHC profiles resulted in three clusters, two that correlated with species and a third that was the most similar to an average male CHC profile.