Increasing levels of dairy cow mortality pose a challenge to the US dairy industry. The industry's current understanding of dairy cow mortality is reliant upon descriptions largely based on producer or veterinary assumptions regarding cause of death without the benefit of detailed postmortem evaluations. A thorough necropsy is a superior tool for establishing a cause of death, except for cases involving euthanasia for traumatic accidents or severe locomotor disorders. Information provided from a necropsy examination would be most valuable if it were categorized and combined with cow health information in a complete postmortem evaluation designed to guide future management decisions. The objective of this study was to describe dairy cow deaths on a Colorado dairy over a 1-yr period and explore classification systems for necropsy findings that might inform management actions aimed at reducing dairy cow mortality. Throughout the study period a thorough necropsy examination was performed on every cow that died. Based upon this examination each death was characterized by a proximate cause (i.e., the most likely immediate cause of the death). Each proximate cause of death was then categorized using 3 alternate schemes founded on generalized etiologic principles and influenced by previous clinical history and treatments. These schemes included the broad categories commonly used for classifying findings within a review of literature related to dairy cow mortality, a diagnostic scheme used within the problem-oriented veterinary medical record, and an analysis focusing on the primary physiologic system derangement for each death. A total of 2,067 cows were enrolled during the study period of which 1,468 cows freshened, 507 cows were sold, and 94 cows died, resulting in a mortality risk of 6.4 deaths per 100 lactations at risk. The distribution of deaths by parity was significantly different from the herd distribution at the end of study with the largest percentage of death present in parity > or =4. Postmortem findings attributable to a specific cause of death were present for all but 4 of the 94 deaths. Assignment of the proximate causes of death to categories within the 3 alternate schemes provided a means for classifying necropsy findings and causes of death with different levels of detail. Creating categories with more selective groupings may provide a means for capturing specifics related to deaths that can be used to guide management decisions.