ROLE OF THE MICROSCOPE 183 



that similar anatomical structures may not occur in very different en- 

 vironments, for a structure may be held on by factors of heredity, if 

 it is not detrimental to the survival of the species. 



In view of these facts, the questions suggest themselves, is it de- 

 sirable to attempt to use anatomical characters in the construction of 

 a rehable key for the identification and classification of wood, and, if 

 so, what methods should be used in the construction of such a key? 



That the former question should be answered in the affirmative is 

 indicated by many considerations. In the first place, the gross or 

 macroscopic characters of wood, such as color, odor, texture, grain, 

 weight, gloss, hardness, strength, resonance, etc., are difficult to de- 

 scribe accurately, and can be easily altered or disguised in manu- 

 factured products. As soon as an attempt is made to gauge these 

 characters by accurate qualitative or quantitative methods, they be- 

 come more cumbersome and difficult to handle than anatomical char- 

 acters, which can be described accurately and rapidly by means of 

 simple drawings and photographs. In the next place, the gross char- 

 acters are often much more variable and unreliable as diagnostic 

 criteria than the more minute internal characters. Furthermore, as 

 the number of woods to be classified increases, the construction of 

 an accurate key, such as is needed for general scientific purposes, and 

 the use of technical experts in certain phases of commercial work, 

 becomes so difficult that all available properties, microscopic and 

 macroscopic, physical and chemical, must be made use of. 



As has been stated previously, the methods used in the construc- 

 tion of keys for distinguishing different woods have been those of 

 '"trial and error." More or less material of each species has been 

 examined, difference between the various species have been recorded 

 and the results tabulated in a key. Owing to the fact that the varia- 

 bility of diagnostic characters has been underestimated, and only a 

 comparatively limited amount of material has been examined, the 

 criteria of one investigator have been found to be more or less unre- 

 liable by the next investigator, who has substituted for them characters 

 that he considered to be more conservative. It is obvious that if this 

 process is continued long enough, each succeeding investigator testing 

 the accuracy of the work of his predecessors, a thoroughly accurate 

 and reliable key may be obtained eventually. The progress made at 

 any time will be largely proportional to the amount of material of 

 each species that is examined. 



