Phenotypic plasticity enables development to produce multiple phenotypes in response to environmental conditions. Plasticity driven variation has been suggested to play a key role in adaptive divergence, and plasticity itself can evolve. However, the interaction of plasticity with the multiple levels involved with adaptive divergence is less understood. For example, sexual dimorphism can contribute to adaptive variation through ecological sexual dimorphism (ESD), but the contribution of plasticity to this phenomenon is unknown. Therefore, to determine the potential contribution of plasticity to ESD, we used the adaptive radiation of Malawi cichlids. Two mouthbrooding species (Labeotropheus fuelleborni and Tropheops “Red Cheek”, hereafter LF and TRC respectively) with differences in foraging tactics underwent foraging experiments using benthic and limnetic treatments while accounting for sex. Plasticity in craniofacial shape and three functionally important traits were measured. Plasticity was shown, but without any sex-based differences in shape. However, for mechanical advantage traits of the mandible sex by diet interactions were found. This suggests that ESD, may be influenced by phenotypic plasticity that diverges between sexes. Given the involvement of the mandible in parental care in cichlids this may indicate that sexual divergence in plasticity may trade-off against maternal care tactics.

Stock cichlids were obtained from the wild and purchased through the pet trade. At the University of Glasgow, they were kept in 100L aquariums and monitored for breeding and egg holding. Cichlid broods were collected for LF and TRC from females at 3 days post fertilisation (dpf). A total of 101 fish from ten broods were collected for TRC, and 115 fish from four broods were collected for LF. Each brood was raised separately in a 1L conical flask with 1-2 drops of methylene blue to prevent fungus growth and the embryos were kept aerated. At around 20 dpf the yolk was nearly fully absorbed, and each brood was moved into a separate 25L tank to feed independently. After a further 4-6 weeks, the broods were divided and moved into larger (~125L) tanks as part of a mixed family group. In total there were four treatment tanks for each species; two for the benthic treatment and two for the limnetic. Each tank contained the same enrichment and used the same water supply. To limit potential effects from density, each family was divided approximately equally into four treatment tanks containing between 22-26 fish.  

Diet treatment experiment

To test the impact of different foraging and biomechanical demands on morphology, treatment groups were fed either limnetic or benthic food based on previous methods. The nutritional content of food was kept similar to limit the possibility of nutritional effects on morphological plasticity. The limnetic treatment, given to half of the groups, consisted of a ground mixture of flake food, algae wafer, and freeze-dried daphnia which was then sprinkled into the water column to elicit suction feeding. The benthic treatment, given to the other half of treatment groups was the same mixture but air dried o to lava rocks. During feeding these rocks were placed at the bottom of the tank to elicit a biting mode of feeding. Each treatment tank was fed twice daily until satiation (morning and afternoon feeds) for approximately 6-7 months until fish were within the size range of a mature cichlid (approximately 4-8 cm standard length (SL)) and sexual dimorphism in colouration and spawning activity had begun. Smaller fish were excluded from downstream analysis as they were difficult to dissect (n = 4). All fish were euthanised following UK Home Office Schedule 1 guidelines, labelled, and fixed in 10% neutral buffered formalin (NBF). Fish were sexed using colouration and venting. Dissection of internal anatomy was also conducted to confirm sex.

Morphometrics

Following fixation, the craniofacial region was dissected to reveal musculature and allow landmarks to be collected for geometric morphometrics. Craniofacial landmarks were selected based on previous work to ensure that they were relevant to the evolution and functional anatomy of cichlids. Fish were secured to a wax dish with a scale and ID tag and photographed laterally from a fixed distance, with their mouth closed, using a mounted Canon EOS 1100D camera (Canon (UK), Surrey).

For landmark digitisation, the tps suite of software was used (available at: http://life.bio.sunysb.edu/ee/rohlf/software.html). Firstly, a tps file linking photographs was created from all photographs using tpsUtil. Prior to digitisation and to reduce intra-observer variability, the ID tags were removed from the images and the photographs were randomised using tpsUtil. Digitisation of landmarks was conducted using tpsDig2 with a scale factor measured for each image. Following digitisation, the tps files were analysed in R version 4.0.4 using the geomorph package unless stated otherwise. Before any analysis could take place, the landmarks were subjected to a Procrustes superimposition that translated, rotated and scaled specimens to a common centroid size using gpagen in the geomorph R package. Procrustes coordinates were then used for all downstream analyses.