198 MALKUS [chap. 4 



cell in the low latitudes of a heated, rotating system, has been studied in 

 experimental (Fultz, 1949) and theoretical (Kuo, 1954) investigations of simple 

 planetary analogs. The foregoing treatment of a limited portion of the real 

 tropical cell suggests another kind of stability which may not be unrelated. 

 We have examined the steady-state coupling between the small-scale motions 

 occurring within a trade section, and the modifications in the large-scale flow 

 observed between its inflow and outflow ends ; the heat source and its convective 

 distribution are found essential in maintaining the mean flow as observed. 



When either scale of motion suffers alteration, the response of the other is 

 significant in determining the nature and amplitude of the system's fluctuations 

 about its average condition. In the mid-latitude westerlies, disturbances of 

 certain critical sizes, drawing upon stored energy, are able to grow in a 

 self-accelerative fashion, and, in so doing often upset the entire structure 

 of the flow, which thereby fluctuates so wildly that an average picture has only 

 a very restricted meaning as a "steady state". In the tropical easterlies, by 

 contrast, the average and daily pictures bear a closer resemblance to one 

 another. The infrequency of intense disturbances in the lower trades, and the 

 reliable presence of trade cumulus clouds, suggest that the interaction between 

 small- and large-scale processes is stable and perhaps contributes to the overall 

 dynamic stability of the tropical circulation. 



In the framework of the present model, therefore, we might inquire what 

 happens if the net heat source is weakened, for example through a diminution 

 of convective precipitation. A stable coupling of the type suggested would 

 prevail if weakening the heat source leads to decreased subsidence. It is well 

 known that subsidence in the trades is one of the most efifective brakes upon 

 convection, and even a slight release of this brake results in vigorous outbursts 

 of shower activity. 



Any such approach is clearly a drastic oversimplification and its formulation 

 here is intended only to be suggestive. No simple causal relation between 

 heating and trade flow is implied by the existing model ; it merely prescribes 

 relations between the two which must on the average be satisfied. To deal 

 with the complete, time-dependent behavior of the motions much larger 

 circulation branches must be considered. The present theory can, on the other 

 hand, suggest what other steady pr quasi-steady configurations are possible 

 and self-consistent for the section. If such configurations may be interpreted 

 as describing the slow fluctuations about the seasonal mean picture, we may 

 profitably inquire whether a given one of these exhibits a structure likely to 

 restore the seasonal mean or lead to even greater departures from it. 



Re-examining equation (42) and temporarily ignoring variations in gS^/u^, 

 we see that the flow may be restored in two ways : either by a stable interaction 

 between the whole forcing function and the motion field, or by internal re- 

 adjustment between the two right-hand terms. In both categories, only a few 

 of the possible rearrangements will give a consistent relation between the 

 resulting wind-profile and stress distribution ; the remainder are, therefore, 

 excluded. Due to inadequate knowledge of the dependence of the two 



