Background: A challenge in efforts to mitigate the global impact of farm origin antibiotic resistance is a failure to understand dissemination patterns of resistance within the farm via animals or environmental flow. This study focuses on diversity of phenotypic antibiotic resistance within niches on the dairy. The goals were to identify niches where diversity is generated and maintained, identify niches connected by phenotype similarity, and identify niches that narrow diversity. The premise was that diversity is generated in niches with high antibiotic use and disseminate through the dairy system with animal movement or environmental flow. Twelve commercial dairy herds were sampled. Farm niches were defined by housing and function. Housing niches included areas with pre-weaned heifers, weaned and bred heifers, early lactation cows, lactating cows, pregnant nonlactating cows, antibiotic treated cows, and cows leaving the herd. Function niches include milk, waste, water, and feeds. For each sampling time, approximately 100 samples were collected across the niches. We used E. coli as our model and antimicrobial susceptibility tested 4 isolates from each sample against 13 antimicrobials. Latent class analysis (LCA) was used to organize isolate-based resistance patterns and clustered using K-means. Risk factors were assessed using multinomial logistic regression. Each farm was visited 3 times. More than 13,000 E. coli isolates were tested for antimicrobial susceptibility patterns. Based on LCA, 24 different patterns were observed. The most common pattern was pan-susceptible (68%) followed by a pattern with only tetracycline resistance (7%). The greatest diversity was observed in pre-weaned and weaned calves followed by treated animals. E. coli from adult cows were pan-susceptible (approaching 90% of isolates). E. coli from soil samples were pan-susceptible except for samples in pre-weaned housing areas. There were farm-specific patterns and between farm differences in diversity. There are niche-specific patterns of resistance that suggest there is little to no dissemination across niches. The between-farm patterns also suggest that resistance traits are farm-specific and driven more by local features rather than shared across spatially distinct farms.