288 PLANT-BREEDING 



From these short descriptions it is easily seen that one 

 of the aims of the method of selection is the grading of all 

 marks in figures. By this means it becomes possible to 

 give extensive comparative surveys of whole groups of newly 

 selected plants or of the strains derived from them and to 

 draw a parallel between the new and the already existing 

 varieties. Without it the testing of the value of many 

 hundreds of new types would evidently be scarcely possible. 



With the same object in view larger collections have 

 been made and separate buildings have been erected in 

 order to exhibit them in such a way that they are easily 

 accessible whenever a new form must be compared with the 

 older types. For the testing is by no means limited to the 

 ears and the kernels, but embraces the whole plant, during 

 the several periods of its development. Dried spec- 

 imens of seedhngs must be at hand for all of the useful varie- 

 ties. The manner in which the culms produce their more 

 or less numerous side halms must be illustrated, also the 

 length and the stiffness of the straw, on which nearly the 

 whole value of some varieties depends. Moreover the varie- 

 ties are not only to be compared with one another, but their 

 constancy must be tested, and for this purpose some spec- 

 imens of previous years must be preserved. In this way 

 pedigrees of dry specimens have been built up, which convey 

 to the visitor an idea of the whole previous history of the 

 race, and contain, besides the proofs of its constancy, the 

 indications of any deviation, which may have occurred in it. 



These collections are to be considered as li\ing expres- 

 sions of the systems described before. The investigator 

 visits them with the plant he wishes to classify, and is guided 

 by their arrangement so that he may find the corresponding 

 prototype in the easiest way. They have this great advantage 

 over the printed systems that all the single quaHties may 

 be taken into consideration at once, without the labo- 



