112 THE AMERICAN BOTANIST 



ly reduced in number, often only one to a flower. By the time 

 we have used these distinctions for breaking up large groups 

 into smaller ones, we find that the plants in each division be- 

 gin to have recognizable resemblances and from this point on 

 the key-maker selects anything that will serve his purpose for 

 further segregations. Among other things, the number and 

 position of the stamens, the number of carpels, the position, 

 of the leaves on the stem, the number of seeds, the color of 

 the flowers, and the shape of the leaves are favorite charac- 

 teristics. 



' The final arrangement of all these groups is as near as a 

 linear sequence as it can be made, but since evolution has pro- 

 duced types of many kinds, such a sequence is but an approxi- 

 mation. It is the custom to place the monocots first and the 

 simplest dicots next to them, though it is now known that 

 monocots as a group are more complex that these latter plants. 

 In deciding what are or are not simple types of plants the 

 flowers are again considered. Those with separate petals or 

 none at all, with many stamens and carpels, with regular 

 flowers and superior ovaries are regarded as simple types, 

 while the highest have united petals, few stamens and carpels, 

 irregular flowers and inferior ovaries. According to this the 

 dandelion is one of the highest types. Consult almost any 

 kind of a key and the dandelion will be found to end the list. 

 The main difference between a scientific key and one 

 that arranges plants according to color, or habitat, or time of 

 flowering, is that the scientific key is more exact and invari- 

 ably groups related plants together. When one traces his 

 plant by a scientific key he steadily moves toward that part of 

 it in which all the plant's allies are found, and is helped in his 

 progress by the opportunity to compare his specimen with its 

 relatives. By so doing, he gains much more than the name 



I 



