The first intensive experiments were carried by Benard in 1900. He experimented on a fluid of thin layer and observed appearance of hexagonal cells when the instability in the form of convection developed. Rayleigh in 1916 developed the theory which found the condition for the instability with two free surface. He showed that the instability would results if the temperature gradient was large enough so that the so called Rayleigh number
Experiments in the early stage were carried out with fluid heated from bottom at the top surface is open to atmosphere. Thus the top surface is free to move and deform. It was later (around 1960) realized that this can lead to another instability mechanism (Benard-Marangoni convection) due to gradient in surface tension. This mechanism coexists with the Rayleigh's mechanism but dominates in thin layer. Most of the findings reported by Benard were actually due this second instability mechanism. The instabilities driven by surface tension decreases as the layer becomes thicker. Later experiments on thermal convection (with or without free upper surface) have obtained convective cells of many form such as rolls, square and hexagons.