to make change for 84 cents from a dollar, 83 cents from a dollar, etc, , each 

 as a separate problem requiring an individual effort of the memory. Many 

 such individuals can recall a vast number of accurate solutions yet are entirely 

 helpless to solve those problems which fall outside their memory and nnay be 

 thrown into confusion by the problem of 83 cents from $1.01. The normal mind 

 devotes little effort to the problem of making change in comparison with this de- 

 fective mind as the normal mind only needs to understand a general concept. 

 Part of the problem of arriving at a fuller understanding of any system involves 

 the mental manipulation of generalized models rather than a memory of unex- 

 plained details. I am sure that each of us recognizes a consciousness of re- 

 duced effort when the conception of a mental model explains, generalizes, and 

 permits us to forget a previously "unexplained" mass of detail. 



Instrumentation, therefore, is most effective when it is capable of pre- 

 senting proper and sufficient information to permit subsequent generalization 

 into models, rather than when it is capable of presenting only a mass of data. 

 The mixed facies of radiolaria, near bottom turbidity, deep bottom ripples, 

 manganese nodules, etc. , are at present only a syndronne of isolated facts in- 

 dicative of our deficiency of understanding of deep sea processes. The under- 

 standing will be aided by instrumentation that can present the proper infornna- 

 tion in such a way that these facts can be fitted into our concept. The differ- 

 ence of presentation of data of the wire sounding machine and the echo sounder 

 is an excellent example. The wire sounding machine was capable of present- 

 ing isolated data which could fit only into the largest and least complete model, 

 with anomalous soundings as irritating and unexplained facts. The echo sound- 

 er can, in effect, presents a seamount, a trench, or a swell as a unit, relative 

 to some intensively studied example, and permit the thought to handle the more 

 complex unit with reduced effort. Understanding comes much more rapidly 

 from these generalities. How much can be done with a vast memory of sound- 

 ings across the ocean, compared with the knowledge that "here the ocean is 

 pimpled with seamounts, here great trenches parallel steep mountainous coasts, 

 and here are great level plains". . The entire sea must eventually become a unit 

 model which can be applied with understanding to other seas of the universe in 

 space and time. Thus the Revelles, Iselins, Eckarts, Munks, and Sverdrups, 

 dealing with complex models as units rather than with details are left free to 

 associate these into even greater concepts for greater understanding. 



Let us look for a moment in more detail at the mode of growth of this 

 mental model. Its advancing fronts are mainly qualitative or descriptive, its 

 central core is quantitative and general. Satellite concepts, some in advance 

 phase and not yet incorporated in the central core, exist in more or less remote 

 areas.* Qualitative data often are sufficient for the development of the front, 

 but for a test of this model it is necessary for quantitative data to be obtained. 

 The multitude of individual processes are then generalized into a principle or 

 mathematical statement which substitutes, as a simpler process, for the com- 

 plex processes in the prototype. When it becomes necessary to push further 

 and utilize the principles discovered for support of further understanding, con- 

 siderable perfection frequently is required, second and higher order terms 

 must be investigated, and these in turn nnust be tested. 



Thus we have in the development of some portion of our understanding. 



* - Some concepts apparently are doomed to an eternal orbit as dead 

 satellites. The study of dinosaur footprints in Connecticut (a complex taxonomy) 

 probably never will be incorporated significantly into our understanding of the 

 Mesozoic. 



