The purpose of this research was to develop an algorithm that would allow the analysis of microarray data from a continuous temporal series of biological data without replicates at any time points. Building upon recent work by Bar-Joseph, Gerber, Gifford, Jaakkola, and Simon, this algorithm represents the characteristic behavior of each cluster by a cubic spline, with variations from this characteristic behavior being governed by cluster and gene-specific parameters. The algorithm converges to a local minimum of the log likelihood function, so is sensitive to how the initial clusters are assigned. Initial assignment is particularly problematic for noisy data. To overcome this difficulty, the data were first smoothed. Additionally, several heuristics were developed to split and/or combine clusters in order to find better solutions. The algorithm was used to analyze data collected throughout the estrous cycle in the rat with the hope of discovering hormone dependent genes that alter mammary development in the virgin animal.