Laboratory science practice relies heavily on procedural knowledge and know-how for successful research, however, its partial tacit nature makes it difficult to capture and codify such knowledge - much of the skill and understanding of laboratory science is learnt and transferred socially. Here we apply the concept of a pattern language, first developed in the domain of architecture by Christopher Alexander, and apply it to the domain of molecular biology as a way to capture and encode more of our laboratory know-how, and further extend this idea to the semantic web. Our traditional forms for organizing and distributing laboratory science procedures rely on journal articles, video, and more formal approaches utilizing workflows and ontologies, yet each of these representations still leaves out some detail in the form of assumptions, background knowledge, or context that is required for the efficient understanding and successful reuse of laboratory knowledge. We show here how laboratory patterns can restore what is missing from our existing descriptions of laboratory procedures. A semantic model is provided that affords new ways to integrate the knowledge captured in laboratory patterns on the Web. A laboratory pattern language offers us a novel and complementary way to capture and communicate more of our knowledge in a networked environment.