Insulation draws water

William B. Rose

Research output: Contribution to journalArticlepeer-review


In the late 1930s, an architect and two researchers created a version of hygrothermal building science for the United States that focused on moisture conditions in exterior materials during cold weather. The version they created was partial, and it was biased: It highlighted the importance of vapor transport, while it obscured the importance of temperature impact. They based their argument on the prevention of "condensation," yet they failed to provide a definition of condensation sufficient for use as a performance measure or criterion. They produced prescriptive recommendations that later became code requirements, and these prescriptions embodied the incomplete and biased nature of their analysis. They supported their argument with a flawed and misleading analogy. They and their followers left a legacy of consumer fear of ill-defined moisture effects in buildings and of designers assigning excessive importance to prescriptive measures. Their version provides inadequate preparation for the anticipated re-insulation of millions of U.S. buildings in the years to come. This paper will provide a short description of the hygrothermal issues involved. It will trace the development of the condensation version by Rogers, Teesdale, and Rowley and the efforts that followed up to 1952. It will explain the legacy and impact of this approach related to existing building re-insulation and professional practice in design and architecture. It will propose a framework for reviewing the link between moisture control prescriptive requirements and performance outcomes.

Original languageEnglish (US)
JournalJournal of Testing and Evaluation
Issue number1
StatePublished - Jan 2011


  • Condensation
  • Insulation
  • Moisture control

ASJC Scopus subject areas

  • General Materials Science
  • Mechanics of Materials
  • Mechanical Engineering


Dive into the research topics of 'Insulation draws water'. Together they form a unique fingerprint.

Cite this