VUl INTRODUCTION. 



I have followed it in the arrangement of the genera, and have in that part 

 been more ftdl than is usual where the species are arranged according to the 

 natural orders. After he has determined the genus, the student will find a 

 direct reference to the species it contains, and I have headed my pages in a 

 manner to make that reference as easy as possible. I here, however, subjoin a 

 slight sketch of De CandoUe's arrangement. 



According to this author, the Vegetable World is separated into two large 

 divisions. In the first, the structure of the plants is partly made up of vessels 

 or pipes ; while in the second it is formed entirely of cells. These divisions nearly 

 coincide with the Phanerogamy and Cryptogamy of the Linnsean system : ex- 

 cept that the Ferns and Miscellanese of that system find their place in the first 

 division of De CandoUe. 



This first division, which is all I have to do with in the present work, com- 

 prises three classes, Dicotyledonous, Monocotyledonous, and Cryptogamous. 



Dicotyledonous plants, besides the character drawn from the seed, which is not 

 always accessible, are distinguished by the branching and anastomosing veins of 

 the leaves, and by the parts of the flower being four or five, or some multiple of 

 these numbers. In woody plants they are more certainly marked by the struc- 

 tm'e of the trunk, which is composed of concentric layers, a new layer being 

 added on the outside of the woody centre each year. 



Monocotyledonous plants, on the contrary, have simple nerves, extending from 

 one end of the leaf to the other, without branching into veins. The parts of the 

 flower are threes or multiples of three; and in the larger and more permanent trunks, 

 the gi'owth takes place internally, and the section exhibits no concentric layers. 



The Ferns and MiscellancBe have no flowers. 



These rules are subject to many exceptions. Berberis, Zeontice, Loranthus, 

 Frankenia, Peplus, and Rumex, have six stamens. The Cruciferse have also six 

 stamens, though only four petals. Several plants of difi"erent families have only 

 a midrib, and there are consequently no anastomosing veins. In Lathyrus Nis- 

 solia the apparent leaf is very much like that of a grass. 



On the other hand, among monocotyledonous plants, Maianthemum, and the 

 Em'opean species of Paris, have four petals and as many stamens. Arum, 

 Calla, Smilax, Tamus, Buscus, Paris, have anastomosing veins. Hydrocharis and 

 the Alismaceee have two sets of nerves, or veins, crossing each other ; one set pro- 

 ceeding from the base of the leaf, and the other from the midrib. In the Pota- 

 meae something of a similar structm-e may be observed ; the principal nerves 

 being united or crossed by veins nearly at right angles. I am not aware that 

 this arrangement is observable in any dicotyledonous plant. 



The Dicotyledons are divided by De Candolle into Thalamiflora, CaVycijlone, 

 CorolliflorcB, and MonocJilamydea. In the first, the stamens ai-e inserted upon a re- 



