The successful use of remote site incubators (RSIs) to rear eggs of Arctic Grayling Thymallus arcticus along Montana streams has sparked interest in reestablishing the species in Michigan. As a preparatory step, we assessed the efficacy of RSIs by deploying them along three Michigan streams during 2 years using surrogate eggs from Rainbow Trout Oncorhynchus mykiss. Our objectives were to (1) compare hatching success between two different RSI designs (19-L versus 265-L RSIs), (2) test whether the removal of dead eggs ("picking") from 19-L RSIs affected hatching success, and (3) develop a simple model to predict fry yield and its uncertainty. Overall survival was 41.3% in 2018 and 52.4% in 2019. Differences in survival between unpicked 19-L and 265-L RSIs tended to be small, with mean differences from 4.82% (95% CI = -0.60 to +10.25) in 2018 to 0.08% (95% CI = -0.14 to +0.30) in 2019. On average, picked 19-L RSIs had greater, although not always statistically significant, survival than unpicked 19-L RSIs during both years (mean difference = 1.6% [2018] and 10.4% [2019]). We documented a significant positive correlation between survival and RSI flow rate. Survival abruptly declined in unpicked 19-L RSIs when RSI flow rates dropped below~0.3 L/min, suggesting that removing dead eggs from 19-L RSIs likely increased survival when RSI flow rates were <0.3 L/min. The most notable result from our fry yield model was that increasing the number of RSIs reduced the coefficient of variation in fry yield following a pattern of diminishing returns, suggesting two or three RSIs usually will be a good choice. We showed that 19-L and 265-L RSIs can be used successfully in Michigan streams, with our model providing a tool for managers to explore the relative importance of several properties of RSI design and operation on fry yield and uncertainty.