Studying parallel evolution allows us to draw conclusions about the repeatability of adaptive evolution. Whereas populations likely experience similar selective pressures in similar environments, it is not clear if this will always result in parallel divergence of ecologically relevant traits. Our study investigates the extent of parallelism associated with the evolution of pelvic spine reduction in brook stickleback populations. We find that populations with parallel divergence in pelvic spine morphology do not exhibit parallel divergence in head and body morphology but do exhibit parallel divergence in diet. In addition, we compare these patterns associated with pelvic reduction in brook stickleback to well-studied patterns of divergence between spined and unspined threespine stickleback. Whereas spine reduction is associated with littoral habitats and a benthic diet in threespine stickleback, spine reduction in brook stickleback is associated with a planktonic diet.  Hence, we find that pelvic spine divergence is associated with largely non-parallel ecological consequences across species.

For the morphological dataset, to quantify two-dimensional body shape, we digitized 27 anatomical landmarks on left-lateral photographs using tpsDIG2w32, version 2.31 (Rohlf, 2018). All landmarks were visible from the lateral side of the fish. The locations of landmarks are shown in Figure in the manuscript text. The landmark data is stored as a TPS file. Other data associated with the samples, including pelvic morphology, source population, sex, size, and year of collection, are in a separate CSV file (see README file).

For the stable isotopes analysis, stickleback, zooplankton, and benthic macroinvertebrate tissues were thawed then rinsed with distilled H₂O to clean off any lake-debris or mud. We then dried each sample at 65°C for 48 hrs in an incubator oven, ground it into a powder in liquid nitrogen using a mortar and pestle, then packed the ground tissue into 4 x 6 mm tin capsules for isotope analysis. Stable isotope analysis was performed on packed capsules using a Carbon and Nitrogen Ratio Mass Spectrophotometer at the University of Calgary Geosciences Isotope Analysis Laboratory. Stable isotope ratios are expressed as a delta notation (δ) which is defined as the parts per million (‰) difference from a universal standard. The standard material for δ13C is Pee Dee belemnite (PDB) limestone, and for δ15N it is atmospheric nitrogen (both ‰ values arbitrarily set at 0 ‰). Stable isotope data and all other data associated with the samples, including pelvic morphology, source population, sex, size, and year of collection, are in a single CSV file (see README file for more details).