Many fish species display inter-population and inter-individual egg size variation. Intra-specific differences in egg size seemingly reflect both energetic experiences of individual spawning fish and long-term population responses to differing ecosystems. Optimal egg size theory implies that selection influences a population's mean egg size in response to its early-life environment, given the well-established trade-off between egg size and fecundity. Currently, there is strong interest in rehabilitation of Laurentian Great Lakes cisco, Coregonus artedi, which is characterized by inter-population variation of morphological and behavioral traits. However, the extent of cisco egg size variation is under-described. In fall 2018 and 2019, we collected egg samples by stripping ripe females at seven total locations in four Great Lakes. We measured unfertilized egg diameters using imaging software and compared mean egg diameters among locations with and without including maternal total length as a covariate. Lake Michigan females produced the largest eggs overall but were excluded from analyses using the total length covariate because of their significantly larger body sizes. Maternal length had a positive effect on egg size, and when accounting for this effect, females in Lake Huron produced the largest eggs followed by Lake Ontario and Lake Superior. We also found that egg size varied among locations within Lake Superior. These findings aligned with observations of morphological and behavioral differences among populations and suggest that cisco phenotypic variation at a fine spatial scale extends to reproductive biology. Consideration of cisco reproductive traits, such as egg size, may inform restoration strategies, including supplemental stocking.

Cisco (Coregonus artedi) and lake whitefish (Coregonus clupeaformis) are native fish species of management concern in the Laurentian Great Lakes that often overlap in spawning locations and timing. Thus, species-level inference from in situ sampling requires methods to differentiate their eggs. Genetic barcoding and hatching eggs to visually identify larvae are used but can be time and cost intensive. Observations in published literature indicate that lake whitefish eggs may be larger than cisco eggs in the Great Lakes, but this has not yet been substantiated. Samples from shared spawning grounds are unlikely to contain similarly sized or colored eggs from other species. Thus, we assessed whether lake whitefish and cisco eggs could be separated based on size alone. Fertilized, hardened eggs were collected in situ during spawning at Elk Rapids, Lake Michigan and Chaumont Bay, Lake Ontario and preserved in ethanol. Individual eggs were measured and genetically identified. Mean diameter for cisco (2.45 mm, SD = 0.22, n = 444) was smaller than for lake whitefish (3.21 mm, SD = 0.20, n = 99). We used classification trees to identify a species-separating size threshold of 2.88 mm (95% bootstrap CI = [2.877, 2.976]), which classified eggs with an accuracy rate of 96%. Differences between species across other samples from the same locations were mostly consistent with the threshold size, but we suggest validation if applying this method to other populations. Separation of cisco and lake whitefish eggs by diameter can be accurate, efficient, and especially suitable for large sample sizes. (C) 2020 The US Geological Survey. Published by Elsevier B.V. on behalf of International Association for Great Lakes Research. All rights reserved.