Piacsek 



Fig. 5 - Va riation of horizontally 

 averaged temperature with depthfor 

 experiment B 



The results for Case 2 are presented in Figs. 6 and 7, for L = 3 cm, D = 

 4.5 cm, 3t/Bz|^ = d = 3.0°C/cm, and water as the working fluid. The perturba- 

 tion was applied only after 2 seconds, and had n = 1, A = .001 °C, and H = 

 1.5 cm. Since this is a much deeper system and has a stable layer forming 

 on the bottom, the transient convection pattern is very different, although the 

 final state is not that much. Only one noticeable period of oscillation was car- 

 ried out by the fluid, and the latter again came to a steady state after -^180 

 seconds. The downward- moving initial blobs had weakened long before they 

 reached the stable layer: they seemed to consist of wide but weak "tongues" 

 with a "finger" growing inside them. Eventually the tongues retracted and 

 formed fingers in the steady state. The maximum depression of the top of the 

 stable layer came when the tongues were already in the retracting stage, indi- 

 cating that the temperature profiles are a poor indicator of the actual fluid mo- 

 tion, for the reasons mentioned in discussing Case 1. A further evidence of 

 this was the almost symmetrical pattern in the streamlines (not shown), indi- 

 cating that the strength of the up- and downward- moving columns is not greatly 

 different. One must bear in mind that the fluid particles continue to gain heat 

 from the time they leave the top until they return to it, so that in this type of 

 convection the temperature field is not at all reliable to assess local circula- 

 tion strengths, though it may be used to study the geometry of cell patterns, as 

 the schlieren photographs have shown. 



Figure 7 shows the vertical variation of the horizontally averaged tempera- 

 ture field, which again shows that in a substantial portion of the flow the tem- 

 perature gradient is reversed, for the same reasons as in Case 1. 



768 



