426 THE SCIENTIFIC MONTHLY 



apparatus needed will be of a standard type that can be readily- 

 obtained, but in most cases it will be necessary to build special ap- 

 paratus. There are two ways to design research apparatus; the first, 

 is to lay out an expensive and elaborate design on the drafting board 

 without experimental trial of any of its parts, the second, is to carry 

 on the experimental work as far as possible with inexpensive, tempor- 

 ary means, until the soundness of the method is assured, and then de- 

 signing the apparatus from the knowledge gained in this manner. The 

 first method may work out satisfactorily but more often it does not, and 

 an expensive instrument or machine must be greatly altered or 

 scrapped. The second is not as rapid nor as convenient, but it is 

 safe and sure and should certainly be used unless there is a large fund 

 of past experience to draw from in regard to that particular subject. 



After the general type of apparatus is decided upon, there is its 

 detailed construction to consider. If it is to be used for only a short 

 time, there is no need to go into elaborate design, or neat finish; the 

 main points to think of are cheapness and quickness of construction 

 and ample opportunities for making alterations. It is always best to 

 rebuild a standard type of machine if possible, and it is surprising how 

 a little ingenuity will make a working device out of the most common 

 parts. It is natural that an experimenter should desire to have an 

 original and neat piece of apparatus, but first it should be determined 

 whether the increased efficiency in operation will warrant the necessary 

 expense. However, it is often the case that a little extra expense if ap- 

 plied efficiently to finishing a piece of apparatus will effect a saving 

 in the end by preserving it from the attacks of moisture or chemical 

 fumes. At the same time a well finished job, even though it will not 

 actually give more accurate results, takes it out of the home made class, 

 and gives to the report illustrated by its photographs, a workmanlike 

 tone that is of the greatest value in raising the standard of that par- 

 ticular investigation. Although an expensive and well finished instru- 

 ment does not assure excellence in the experimental work, yet the per- 

 son reading the report is unconsciously affected by the appearance of 

 the apparatus, crude apparatus being associated with uncertain results. 

 On the other hand, there is undoubtedly a tremendous amount of money 

 spent each year on unnecessarily elaborate machines and instruments, 

 but the waste is usually due more to the fact that the complete design is 

 unworkable, than that the elaboration is unnecessary. 



There are other types of research that extend indefinitely into the 

 future, using principally the same apparatus in all the tests. Examples 

 of this are tests of material strength, or model testing is towing basins 

 or wind tunnels. In these cases the conditions are far different from 

 the isolated investigation, and it is simply a question of how far it will 

 be advisable to go, in using automatic and recording devices to reduce 



